h3-blackwell-runtime/benchmarks/gb10-post-fc2-warmed-step-stats_cuda_kern_exec_trace.csv at fl2va-direct-runtime

Daniel Maddern d345aae6e9 Record post-FC2 production profile

2026-08-26 16:14:12 +07:00

940 KiB

Raw Permalink Blame History

1	API Start (ns)	API Dur (ns)	Queue Start (ns)	Queue Dur (ns)	Kernel Start (ns)	Kernel Dur (ns)	Total Dur (ns)	PID	TID	API Function	GridXYZ	BlockXYZ	Kernel Name
2	7267792	13408			7275376	2144	13408	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
3	7290064	5728			7292688	2816	5728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)
4	7303328	5504			7305872	3392	5936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
5	7312624	3072			7313232	1088	3072	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
6	7320240	3056			7320496	1088	3056	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
7	7326912	2624			7326704	1056	2624	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)
8	7332464	2944			7332624	2080	2944	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
9	7339552	2752			7339504	1824	2752	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
10	7354064	6960			7357872	2752	6960	73	73	cudaLaunchKernel	13 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
11	7368368	6496			7370736	5728	8096	73	73	cudaLaunchKernel	26 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl<at::native::BinaryFunctor<c10::BFloat16, c10::BFloat16, c10::BFloat16, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
12	7381248	2672			7381104	1120	2672	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
13	7392128	4000			7392976	1984	4000	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
14	7416896	3616			7417616	38432	39152	73	73	cudaLaunchKernel	3497 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
15	7457216	5488			7459632	48832	51248	73	73	cudaLaunchKernel	6993 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)4, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 12)]::operator ()() const::[lambda(c10::BFloat16) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
16	7471440	6608	7478048	31952	7510000	71520	110080	73	73	cudaLaunchKernel	6993 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
17	7645392	5456			7647600	4915936	4918144	73	73	cuLaunchKernel	2336 3 1	256 1 1	void cutlass::Kernel2<cutlass_80_simt_sgemm_256x128_8x4_tn_align1>(T1::Params)
18	7660272	3168	7663440	4902048	12565488	5035136	9940352	73	73	cudaLaunchKernel	195804 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
19	7680208	3104	7683312	9918336	17601648	3808	9925248	73	73	cudaLaunchKernel	26 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
20	7749040	5552	7754592	9853168	17607760	44096	9902816	73	73	cuLaunchKernel	56 6 1	128 1 1	void cutlass::Kernel2<cutlass_80_simt_sgemm_128x64_8x5_tt_align1>(T1::Params)
21	7760960	2784	7763744	9890000	17653744	30368	9923152	73	73	cudaLaunchKernel	2174 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
22	7789584	14368	7803952	9881536	17685488	3744	9899648	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
23	7808960	2624	7811584	9880144	17691728	1792	9884560	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
24	7816096	3040	7819136	9876944	17696080	1344	9881328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)
25	7822960	3856	7826816	9873264	17700080	2048	9879168	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
26	7830640	4176	7834816	9869456	17704272	1536	9875168	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
27	7838288	3408	7841696	9866672	17708368	1280	9871360	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
28	7846032	2512	7848544	9863824	17712368	1376	9867712	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnSelf_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)
29	7851344	3104	7854448	9862112	17716560	2080	9867296	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::DivFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
30	7884416	5552	7889968	9830432	17720400	2848	9838832	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnOther_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)
31	17749936	2864			17749968	1024	2864	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::CUDAFunctorOnOther_add<float>, std::array<char *, (unsigned long)2>>(int, T2, T3)
32	17804720	6544			17807664	3766816	3769760	73	73	cudaLaunchKernel	1536 3 1	128 1 1	void at::native::<unnamed>::CatArrayBatchedCopy_alignedK_contig<at::native::<unnamed>::OpaqueType<(unsigned int)2>, unsigned int, (int)2, (int)128, (int)1, (int)8>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)
33	26357072	19776			26372208	8192	23328	73	73	cudaLaunchKernel	222 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
34	26411856	6864			26413968	27680	29792	73	73	cudaLaunchKernel	7090 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
35	26448080	6096			26449776	52384	54080	73	73	cudaLaunchKernel	96 3 1	512 1 1	void at::native::<unnamed>::CatArrayBatchedCopy<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)2, (int)64, (int)64>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)
36	26467520	8432	26475952	27200	26503152	31872	67504	73	73	cudaLaunchKernel	1773 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::cos_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
37	26485712	7200	26492912	43104	26536016	39456	89760	73	73	cudaLaunchKernel	1773 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::sin_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
38	26503344	7696	26511040	66864	26577904	39712	114272	73	73	cudaLaunchKernel	1773 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::neg_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
39	26524576	5744	26530320	88544	26618864	873696	967984	73	73	cudaLaunchKernel	1536 4 1	128 1 1	void at::native::<unnamed>::CatArrayBatchedCopy_alignedK_contig<at::native::<unnamed>::OpaqueType<(unsigned int)4>, unsigned int, (int)3, (int)128, (int)1, (int)16>(T1 *, at::native::<unnamed>::CatArrInputTensorMetadata<T1, T2, T4, T5>, at::native::<unnamed>::TensorSizeStride<T2, (unsigned int)4>, int, T2)
40	26542304	3824	26546128	949280	27495408	217120	1170224	73	73	cudaLaunchKernel	7090 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
41	26580608	5376	26585984	1128560	27714544	2240	1136176	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
42	26596912	4336	26601248	1117488	27718736	1536	1123360	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
43	26610752	7728	26618480	1104256	27722736	3424	1115408	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
44	26623808	10192	26634000	1094848	27728848	3168	1108208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
45	26644832	6224	26651056	1082016	27733072	3136	1091376	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
46	26750464	31104	26781568	955920	27737488	4736	991760	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
47	26817168	16720	26833888	909616	27743504	1088	927424	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
48	26842128	5984	26848112	899456	27747568	1632	907072	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
49	26881584	13776	26895360	856176	27751536	1504	871456	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
50	26933984	31456	26965440	790320	27755760	2016	823792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
51	27338800	37296	27376096	383760	27759856	30048	451104	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
52	27606992	42304	27649296	143072	27792368	3544672	3730048	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
53	27744576	4640	27749216	3589264	31338480	2464	3596368	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
54	27753472	3216	27756688	3585888	31342576	1984	3591088	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
55	27834368	7616	27841984	3504688	31346672	49568	3561872	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
56	27856400	16064	27872464	3526432	31398896	3452992	6995488	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
57	27875376	14016	27889392	6964448	34853840	2912	6981376	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
58	27895376	18464	27913840	6944128	34857968	2466752	9429344	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
59	28438432	12560	28450992	8875072	37326064	1920	8889552	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
60	28955424	19216	28974640	8355520	37330160	24059776	32434512	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
61	29491888	43168	29535056	31856160	61391216	11842304	43741632	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
62	29793040	28320	29821360	43430944	73252304	2395136	45854400	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
63	29899920	23168	29923088	45726176	75649264	4001984	49751328	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
64	29971760	19104	29990864	49663008	79653872	3999008	53681120	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
65	30031744	16336	30048080	53606784	83654864	5297792	58920912	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
66	30084848	13088	30097936	58857152	88955088	5731328	64601568	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
67	30238096	12496	30250592	64438928	94689520	240315488	304766912	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
68	30402032	15632	30417664	304589328	335006992	2176224	306781184	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
69	30424080	12784	30436864	306748144	337185008	1952	306762880	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
70	30460496	9984	30470480	306718624	337189104	1376	306729984	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
71	30491344	5984	30497328	306695872	337193200	1504	306703360	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
72	30616272	10752	30627024	306570272	337197296	40608	306621632	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
73	30657360	27824	30685184	306555088	337240272	2775488	309358400	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
74	30757632	9776	30767408	309249984	340017392	1888	309261648	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
75	30860752	10288	30871040	309150448	340021488	7702240	316862976	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
76	31210832	13904	31224736	316500304	347725040	5461056	321975264	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
77	31254736	8192	31262928	321925152	353188080	3717632	325650976	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
78	31305920	8032	31313952	325593296	356907248	5856	325607184	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
79	31319328	5072	31324400	325591040	356915440	3776	325599888	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
80	31353824	3712	31357536	325564048	356921584	60928	325628688	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
81	31360320	3296	31363616	325620432	356984048	4091840	329715568	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
82	31364384	3520	31367904	329710064	361077968	1952	329715536	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
83	31368512	3872	31372384	329709712	361082096	2437024	332150608	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
84	31411328	3808	31415136	332105296	363520432	1920	332111024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
85	31464768	3056	31467824	332056512	363524336	34563520	366623088	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
86	31511392	3584	31514976	366574512	398089488	140224	366718320	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
87	31517360	5072	31522432	366709360	398231792	11923520	378637952	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
88	31522864	2624	31525488	378631808	410157296	2464	378636896	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
89	31526352	4816	31531168	378630224	410161392	10756480	389391520	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
90	31555888	2864	31558752	389361808	420920560	1920	389366592	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
91	31570192	4048	31574240	389350416	420924656	3040	389357504	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
92	31658144	4288	31662432	389267344	420929776	16202336	405473968	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
93	31696912	2832	31699744	405433840	437133584	5455232	410891904	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
94	31715264	4176	31719440	410871008	442590448	1984	410877168	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
95	31725792	3056	31728848	410865696	442594544	1440	410870192	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
96	31733712	3488	31737200	410861440	442598640	1376	410866304	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
97	31742144	3552	31745696	410857040	442602736	3168	410863760	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
98	31751328	3232	31754560	410854288	442608848	1344	410858864	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
99	31768288	4032	31772320	410840720	442613040	4608	410849360	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
100	31778848	3312	31782160	410836832	442618992	1120	410841264	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
101	31786592	2992	31789584	410833632	442623216	1632	410838256	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
102	31795744	2992	31798736	410828576	442627312	1472	410833040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
103	31806144	3264	31809408	410822000	442631408	1248	410826512	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
104	31882768	3856	31886624	410748880	442635504	29472	410782208	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
105	31913904	2912	31916816	410749440	442666256	3479584	414231936	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
106	31933664	2976	31936640	414212208	446148848	2368	414217552	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
107	31940592	2832	31943424	414209520	446152944	2112	414214464	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
108	31960448	2640	31963088	414193952	446157040	48416	414245008	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
109	31965232	2880	31968112	414240128	446208240	3449888	417692896	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
110	31968512	2416	31970928	417689216	449660144	4032	417695664	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
111	31971488	2752	31974240	417692048	449666288	2435424	420130224	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
112	32013936	2976	32016912	420087520	452104432	1888	420092384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
113	32061952	2944	32064896	420043632	452108528	25770432	445817008	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
114	32179824	4736	32184560	445697024	477881584	11819968	457521728	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
115	32218688	3456	32222144	457482480	489704624	2346688	459832624	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
116	32238848	3584	32242432	459811312	492053744	3809824	463624720	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
117	32246832	2832	32249664	463615408	495865072	3947776	467566016	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
118	32258208	3104	32261312	467554352	499815664	5406624	472964080	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
119	32269968	3104	32273072	472951360	505224432	5664096	478618560	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
120	32286592	3648	32290240	478601008	510891248	240444576	719049232	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
121	32323632	3488	32327120	719011648	751338768	2172448	721187584	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
122	32327984	3088	32331072	721181584	753512656	1952	721186624	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
123	32338128	3184	32341312	721175472	753516784	2912	721181568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
124	32346992	3056	32350048	721170928	753520976	1536	721175520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
125	32372752	2944	32375696	721149312	753525008	46304	721198560	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
126	32388960	3136	32392096	721180976	753573072	2749568	723933680	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
127	32420640	3104	32423744	723900848	756324592	1920	723905872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
128	32463184	3280	32466464	723862192	756328656	7743936	731609408	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
129	32507568	2832	32510400	731563888	764074288	5461696	737028416	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
130	32524576	3072	32527648	737010608	769538256	3462944	740476624	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
131	32542416	3168	32545584	740456896	773002480	2304	740462368	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
132	32549600	2704	32552304	740454272	773006576	1984	740458960	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
133	32568224	2784	32571008	740439664	773010672	46624	740489072	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
134	32572784	2864	32575648	740483152	773058800	3966080	744452096	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
135	32576096	2512	32578608	744449216	777027824	1952	744453680	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
136	32579200	2528	32581728	744450192	777031920	2729952	747182672	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
137	32612736	2864	32615600	747148352	779763952	1856	747153072	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
138	32654352	3184	32657536	747110512	779768048	33135360	780249056	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
139	32685936	2784	32688720	780216992	812905712	186080	780405856	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
140	32690304	3744	32694048	780400048	813094096	12695904	793099696	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
141	32694496	2816	32697312	793094416	825791728	2176	793099408	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
142	32697712	2880	32700592	793095328	825795920	10475328	803573536	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
143	32715072	2736	32717808	803554720	836272528	2176	803559632	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
144	32724256	2976	32727232	803549232	836276464	832	803553040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
145	32758416	2864	32761280	803517072	836278352	15614944	819134880	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
146	32788016	2768	32790784	819103792	851894576	5572000	824678560	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
147	32803040	3536	32806576	824662592	857469168	2336	824668464	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
148	32812352	3120	32815472	824657792	857473264	3104	824664016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
149	32819408	2592	32822000	824657408	857479408	3264	824663264	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
150	32826160	2928	32829088	824656464	857485552	11680	824671072	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
151	32833712	3520	32837232	824662624	857499856	6144	824672288	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
152	32853184	3360	32856544	824652816	857509360	5600	824661776	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
153	32861760	3376	32865136	824651136	857516272	2368	824656880	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
154	32869712	3088	32872800	824647568	857520368	3360	824654016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
155	32877088	2624	32879712	824646800	857526512	2944	824652368	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
156	32885056	3264	32888320	824642416	857530736	8192	824653872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
157	32952736	3216	32955952	824584896	857540848	56608	824644720	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
158	32981504	2896	32984400	824615840	857600240	3673856	828292592	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
159	32999120	2944	33002064	828273312	861275376	2464	828278720	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
160	33006048	2736	33008784	828270656	861279440	2048	828275440	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
161	33024304	2640	33026944	828256592	861283536	48960	828308192	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
162	33028848	2640	33031488	828303248	861334736	3813056	832118944	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
163	33031888	2800	33034688	832115472	865150160	3776	832122048	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
164	33035216	2496	33037712	832118592	865156304	2434944	834556032	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
165	33080784	2976	33083760	834509696	867593456	1888	834514560	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
166	33125728	3296	33129024	834468560	867597584	23937632	858409488	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
167	33213920	4080	33218000	858319648	891537648	12480864	870804592	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
168	33245472	3360	33248832	870772624	904021456	2345600	873121584	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
169	33263808	3408	33267216	873102048	906369264	3709440	876814896	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
170	33270080	2816	33272896	876807312	910080208	3724512	880534640	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
171	33280032	2736	33282768	880524800	913807568	5534656	886062192	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
172	33290304	2768	33293072	886050432	919343504	5795488	891848688	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
173	33303040	3056	33306096	891835136	925141232	239349632	1131187824	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
174	33336304	3424	33339728	1131153472	1164493200	2170784	1133327680	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
175	33340192	4048	33344240	1133321696	1166665936	1920	1133327664	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
176	33350944	2816	33353760	1133316304	1166670064	2912	1133322032	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
177	33358672	3152	33361824	1133312464	1166674288	1504	1133317120	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
178	33381392	2912	33384304	1133293984	1166678288	47104	1133344000	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
179	33396976	3360	33400336	1133328064	1166728400	2757664	1136089088	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
180	33428064	2976	33431040	1136058096	1169489136	1952	1136063024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
181	33468256	3136	33471392	1136021840	1169493232	7799744	1143824720	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
182	33509632	3280	33512912	1143782208	1177295120	5451936	1149237424	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
183	33523600	3088	33526688	1149222192	1182748880	3475680	1152700960	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
184	33539920	3248	33543168	1152684272	1186227440	2400	1152689920	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
185	33547248	3136	33550384	1152681120	1186231504	2080	1152686336	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
186	33565648	2736	33568384	1152667216	1186235600	46016	1152715968	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
187	33570064	2768	33572832	1152710064	1186282896	3462240	1156175072	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
188	33573280	2640	33575920	1156171008	1189746928	1984	1156175632	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
189	33576480	2656	33579136	1156171856	1189750992	2452064	1158626576	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
190	33616560	2800	33619360	1158585168	1192204528	1888	1158589856	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
191	33656784	2640	33659424	1158549200	1192208624	34645696	1193197536	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
192	33687520	2704	33690224	1193166560	1226856784	139968	1193309232	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
193	33691632	3184	33694816	1193304208	1226999024	12761440	1206068832	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
194	33695344	2784	33698128	1206065056	1239763184	1920	1206069760	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
195	33698672	3184	33701856	1206065424	1239767280	10456096	1216524704	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
196	33715008	2704	33717712	1216507680	1250225392	1952	1216512336	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
197	33724112	2832	33726944	1216502512	1250229456	864	1216506208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
198	33752400	2656	33755056	1216476576	1250231632	15792128	1232271360	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
199	33780528	2448	33782976	1232243792	1266026768	5824160	1238070400	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
200	33794848	3312	33798160	1238054112	1271852272	1952	1238059376	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
201	33803584	2864	33806448	1238049920	1271856368	1408	1238054192	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
202	33810304	2768	33813072	1238047360	1271860432	2848	1238052976	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
203	33817328	3200	33820528	1238044032	1271864560	3200	1238050432	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
204	33824592	3104	33827696	1238043008	1271870704	1376	1238047488	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
205	33836384	3024	33839408	1238035392	1271874800	4416	1238042832	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
206	33844304	2688	33846992	1238033920	1271880912	1120	1238037728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
207	33851264	2640	33853904	1238031136	1271885040	1600	1238035376	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
208	33858224	2944	33861168	1238027968	1271889136	1472	1238032384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
209	33866544	3312	33869856	1238023344	1271893200	1280	1238027936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
210	33931712	2992	33934704	1237962688	1271897392	26016	1237991696	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
211	33958256	2912	33961168	1237964800	1271925968	3545984	1241513696	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
212	33976368	2912	33979280	1241493952	1275473232	2496	1241499360	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
213	33983168	2576	33985744	1241491488	1275477232	2112	1241496176	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
214	34001264	3040	34004304	1241477024	1275481328	47552	1241527616	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
215	34006352	2992	34009344	1241521424	1275530768	4259520	1245783936	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
216	34009760	2672	34012432	1245780992	1279793424	3712	1245787376	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
217	34012944	2608	34015552	1245783984	1279799536	2444704	1248231296	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
218	34050960	3472	34054432	1248191440	1282245872	1824	1248196736	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
219	34103904	2752	34106656	1248143280	1282249936	24882560	1273028592	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
220	34187600	3696	34191296	1272943952	1307135248	12475168	1285422816	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
221	34216080	3328	34219408	1285393248	1319612656	2339296	1287735872	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
222	34233888	2992	34236880	1287717664	1321954544	3722080	1291442736	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
223	34239456	2528	34241984	1291436848	1325678832	3736032	1295175408	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
224	34248800	2864	34251664	1295164768	1329416432	5150368	1300318000	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
225	34259200	2592	34261792	1300307408	1334569200	5873152	1306183152	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
226	34270608	2880	34273488	1306170400	1340443888	238904576	1545077856	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
227	34302512	3696	34306208	1545045104	1579351312	2158912	1547207712	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
228	34306752	2560	34309312	1547203600	1581512912	1984	1547208144	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
229	34315600	2880	34318480	1547198560	1581517040	2880	1547204320	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
230	34323536	9296	34332832	1547188368	1581521200	1536	1547199200	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
231	34351312	2768	34354080	1547171056	1581525136	54208	1547228032	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
232	34366992	3264	34370256	1547210368	1581580624	2747456	1549961088	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
233	34396800	2800	34399600	1549930336	1584329936	1920	1549935056	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
234	34436816	2784	34439600	1549894464	1584334064	7804864	1557702112	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
235	34475856	2848	34478704	1557662368	1592141072	5612512	1563277728	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
236	34489264	3056	34492320	1563263312	1597755632	3661824	1566928192	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
237	34505120	3200	34508320	1566911184	1601419504	2400	1566916784	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
238	34512384	2848	34515232	1566908336	1601423568	2112	1566913296	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
239	34529744	2720	34532464	1566895232	1601427696	46240	1566944192	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
240	34534256	3040	34537296	1566938528	1601475824	3859104	1570800672	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
241	34537760	2592	34540352	1570796976	1605337328	1952	1570801520	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
242	34540848	2416	34543264	1570798192	1605341456	2438848	1573239456	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
243	34571760	2848	34574608	1573208032	1607782640	1920	1573212800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
244	34610608	2720	34613328	1573173408	1607786736	34043744	1607219872	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
245	34641248	2688	34643936	1607187888	1641831824	139840	1607330416	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
246	34645392	2848	34648240	1607324704	1641972944	12366720	1619694272	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
247	34648656	2384	34651040	1619690832	1654341872	1920	1619695136	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
248	34651472	2992	34654464	1619691504	1654345968	10667840	1630362336	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
249	34667152	2928	34670080	1630345968	1665016048	1984	1630350880	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
250	34676304	3024	34679328	1630340816	1665020144	2560	1630346400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
251	34702592	2608	34705200	1630319040	1665024240	15704512	1646026160	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
252	34729936	2912	34732848	1645997504	1680730352	5454720	1651455136	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
253	34743968	3360	34747328	1651439152	1686186480	1952	1651444464	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
254	34752832	2992	34755824	1651433920	1686189744	1440	1651438352	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
255	34759696	2672	34762368	1651431024	1686193392	1312	1651435008	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
256	34766368	3120	34769488	1651428000	1686197488	3264	1651434384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
257	34773616	2736	34776352	1651427248	1686203600	1344	1651431328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
258	34784832	3008	34787840	1651419888	1686207728	4640	1651427536	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
259	34792688	2544	34795232	1651418640	1686213872	1088	1651422272	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
260	34799952	2784	34802736	1651415232	1686217968	1632	1651419648	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
261	34806848	2704	34809552	1651412608	1686222160	1504	1651416816	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
262	34814544	2896	34817440	1651408720	1686226160	1280	1651412896	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
263	34878832	3136	34881968	1651348288	1686230256	27264	1651378688	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
264	34907072	3184	34910256	1651348672	1686258928	3493056	1654844912	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
265	34928800	3280	34932080	1654821792	1689753872	2464	1654827536	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
266	34936064	2816	34938880	1654819056	1689757936	2016	1654823888	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
267	34954144	2720	34956864	1654805168	1689762032	47456	1654855344	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
268	34958784	2800	34961584	1654849568	1689811152	3918208	1658770576	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
269	34962064	2448	34964512	1658766544	1693731056	3392	1658772384	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
270	34964976	2544	34967520	1658769680	1693737200	2845728	1661617952	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
271	35000496	3072	35003568	1661582400	1696585968	1856	1661587328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
272	35048800	6928	35055728	1661534368	1696590096	25301536	1686842832	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
273	35154928	5136	35160064	1686734032	1721894096	11901664	1698640832	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
274	35184208	5216	35189424	1698608960	1733798384	2551328	1701165504	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
275	35203680	2864	35206544	1701145440	1736351984	3738560	1704886864	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
276	35208960	2960	35211920	1704880704	1740092624	3951712	1708835376	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
277	35218672	4384	35223056	1708823264	1744046320	5119744	1713947392	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
278	35230784	5216	35236000	1713931344	1749167344	5697824	1719634384	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
279	35244976	3472	35248448	1719619472	1754867920	240193088	1959816032	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
280	35284608	3456	35288064	1959774480	1995062544	2167328	1961945264	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
281	35288640	2688	35291328	1961941040	1997232368	1920	1961945648	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
282	35298272	3152	35301424	1961935008	1997236432	2944	1961941104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
283	35306416	4000	35310416	1961930240	1997240656	1504	1961935744	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
284	35329728	3936	35333664	1961910960	1997244624	47296	1961962192	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
285	35346480	3856	35350336	1961943472	1997293808	2750016	1964697344	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
286	35377088	3360	35380448	1964665840	2000046288	1952	1964671152	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
287	35421120	4240	35425360	1964625056	2000050416	7806304	1972435600	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
288	35465136	4736	35469872	1972389600	2007859472	5472288	1977866624	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
289	35480464	3872	35484336	1977850432	2013334768	3708736	1981563040	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
290	35498128	3440	35501568	1981544144	2017045712	6624	1981554208	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
291	35505536	3952	35509488	1981544448	2017053936	2144	1981550544	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
292	35524032	3344	35527376	1981530624	2017058000	47936	1981581904	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
293	35529584	3696	35533280	1981573936	2017107216	4063680	1985641312	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
294	35533744	4272	35538016	1985635472	2021173488	1952	1985641696	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
295	35538640	4144	35542784	1985634800	2021177584	2438432	1988077376	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
296	35575616	3312	35578928	1988038848	2023617776	1920	1988044080	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
297	35615008	3456	35618464	1988003408	2023621872	34049568	2022056432	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
298	35655808	3200	35659008	2022014960	2057673968	139872	2022158032	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
299	35660464	3520	35663984	2022152288	2057816272	11888096	2034043904	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
300	35664912	6000	35670912	2034035024	2069705936	1984	2034043008	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
301	35671344	3808	35675152	2034034912	2069710064	10741088	2044779808	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
302	35689952	3264	35693216	2044760656	2080453872	1952	2044765872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
303	35698960	3104	35702064	2044755904	2080457968	2496	2044761504	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
304	35726064	3472	35729536	2044732528	2080462064	16236448	2060972448	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
305	35754320	3344	35757664	2060942960	2096700624	5467584	2066413888	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
306	35771184	4720	35775904	2066394928	2102170832	1952	2066401600	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
307	35782272	3152	35785424	2066389536	2102174960	1440	2066394128	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
308	35789024	2768	35791792	2066387264	2102179056	1312	2066391344	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
309	35795968	2832	35798800	2066383104	2102181904	3456	2066389392	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
310	35803104	3072	35806176	2066381072	2102187248	1344	2066385488	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
311	35815392	3456	35818848	2066372432	2102191280	4608	2066380496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
312	35823776	2896	35826672	2066370816	2102197488	1088	2066374800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
313	35831184	2976	35834160	2066367360	2102201520	1632	2066371968	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
314	35838464	11024	35849488	2066356192	2102205680	1728	2066368944	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
315	35856272	3136	35859408	2066350368	2102209776	1280	2066354784	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
316	35920992	3344	35924336	2066289504	2102213840	26368	2066319216	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
317	35950704	3280	35953984	2066288560	2102242544	3495232	2069787072	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
318	35970352	3616	35973968	2069766560	2105740528	2496	2069772672	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
319	35978512	5664	35984176	2069760448	2105744624	1984	2069768096	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
320	35999808	3184	36002992	2069745728	2105748720	47232	2069796144	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
321	36005040	3312	36008352	2069789488	2105797840	3515456	2073308256	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
322	36009264	4784	36014048	2073302256	2109316304	3360	2073310400	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
323	36015072	4448	36019520	2073302960	2109322480	2434016	2075741424	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
324	36053872	3184	36057056	2075701520	2111758576	1888	2075706592	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
325	36107056	3776	36110832	2075651840	2111762672	25849184	2101504800	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
326	36190640	5216	36195856	2101418752	2137614608	11818912	2113242880	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
327	36218672	3776	36222448	2113214272	2149436720	2354816	2115572864	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
328	36235920	3504	36239424	2115553488	2151792912	3750336	2119307328	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
329	36241808	3184	36244992	2119299792	2155544784	4064384	2123367360	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
330	36251648	4896	36256544	2123355568	2159612112	5416480	2128776944	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
331	36264336	3504	36267840	2128763312	2165031152	5667744	2134434560	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
332	36276400	3280	36279680	2134420496	2170700176	242080160	2376503936	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
333	36308848	4816	36313664	2376469200	2412782864	2159488	2378633504	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
334	36314336	4656	36318992	2378625504	2414944496	1952	2378632112	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
335	36328336	3696	36332032	2378616560	2414948592	2944	2378623200	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
336	36337120	3536	36340656	2378611904	2414952560	1504	2378616944	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
337	36358736	4160	36362896	2378592416	2414955312	49536	2378646112	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
338	36375328	3776	36379104	2378627824	2415006928	2752224	2381383824	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
339	36407392	4912	36412304	2381348192	2417760496	1888	2381354992	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
340	36451232	3680	36454912	2381309648	2417764560	7814080	2389127408	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
341	36492672	3088	36495760	2389085056	2425580816	5456192	2394544336	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
342	36507072	18896	36525968	2394512704	2431038672	3483520	2398015120	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
343	36542272	3568	36545840	2397978784	2434524624	6336	2397988688	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
344	36549792	3424	36553216	2397979376	2434532592	36384	2398019184	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
345	36567776	3184	36570960	2398000544	2434571504	54208	2398057936	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
346	36572704	3856	36576560	2398052288	2434628848	4283872	2402340016	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
347	36577008	3680	36580688	2402333568	2438914256	1952	2402339200	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
348	36581296	2848	36584144	2402334240	2438918384	2440160	2404777248	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
349	36620784	6128	36626912	2404733712	2441360624	1888	2404741728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
350	36668480	3504	36671984	2404692736	2441364720	33763776	2438460016	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
351	36703360	3552	36706912	2438424336	2475131248	139488	2438567376	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
352	36708528	3024	36711552	2438560880	2475272432	12292416	2450856320	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
353	36712096	2912	36715008	2450852560	2487567568	2176	2450857648	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
354	36715456	3760	36719216	2450852512	2487571728	10786304	2461642576	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
355	36733264	3184	36736448	2461624112	2498360560	1920	2461629216	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
356	36742272	8800	36751072	2461613584	2498364656	832	2461623216	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
357	36777024	3056	36780080	2461586688	2498366768	15820672	2477410416	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
358	36803984	3600	36807584	2477382992	2514190576	5490432	2482877024	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
359	36820768	3200	36823968	2482859376	2519683344	2272	2482864848	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
360	36829504	4096	36833600	2482853840	2519687440	1408	2482859344	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
361	36837408	3936	36841344	2482850160	2519691504	3712	2482857808	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
362	36845664	3024	36848688	2482848960	2519697648	8064	2482860048	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
363	36852736	4016	36856752	2482851104	2519707856	1568	2482856688	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
364	36865424	3648	36869072	2482842848	2519711920	8096	2482854592	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
365	36873712	2848	36876560	2482845664	2519722224	6880	2482855392	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
366	36880880	3072	36883952	2482846432	2519730384	1632	2482851136	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
367	36888176	3216	36891392	2482843120	2519734512	14048	2482860384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
368	36896080	3040	36899120	2482851776	2519750896	3648	2482858464	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
369	36960640	3008	36963648	2482793392	2519757040	93376	2482889776	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
370	36987088	3056	36990144	2482863120	2519853264	3592384	2486458560	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
371	37012352	4400	37016752	2486431808	2523448560	2912	2486439120	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
372	37020624	3760	37024384	2486428336	2523452720	2240	2486434336	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
373	37039008	4400	37043408	2486413344	2523456752	47712	2486465456	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
374	37045504	3424	37048928	2486456976	2523505904	3622240	2490082640	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
375	37049360	2816	37052176	2490077632	2527129808	4000	2490084448	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
376	37052672	2880	37055552	2490080432	2527135984	2437120	2492520432	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
377	37098880	4256	37103136	2492473072	2529576208	1856	2492479184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
378	37145296	3792	37149088	2492431184	2529580272	25405312	2517840288	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
379	37232240	4576	37236816	2517750048	2554986864	12071296	2529825920	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
380	37259696	4160	37263856	2529798144	2567062000	2347552	2532149856	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
381	37277456	3328	37280784	2532130048	2569410832	3711616	2535844992	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
382	37283104	2512	37285616	2535838208	2573123824	4098912	2539939632	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
383	37292432	3168	37295600	2539928384	2577223984	5356480	2545288032	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
384	37302816	2704	37305520	2545276992	2582582512	5691456	2550971152	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
385	37313312	3648	37316960	2550960048	2588277008	239531264	2790494960	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
386	37344224	3120	37347344	2790462720	2827810064	2158528	2792624368	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
387	37347904	2576	37350480	2792620160	2829970640	1984	2792624720	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
388	37356720	3328	37360048	2792614720	2829974768	2720	2792620768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
389	37364784	4608	37369392	2792609472	2829978864	1504	2792615584	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
390	37393408	2928	37396336	2792586592	2829982928	46080	2792635600	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
391	37409136	3120	37412256	2792619824	2830032080	2749120	2795372064	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
392	37439296	2944	37442240	2795341360	2832783600	1920	2795346224	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
393	37479072	2704	37481776	2795305920	2832787696	7819616	2803128240	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
394	37516160	3232	37519392	2803090640	2840610032	5449728	2808543600	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
395	37532224	3216	37535440	2808527392	2846062832	3479168	2812009776	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
396	37549472	3072	37552544	2811990864	2849543408	2336	2811996272	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
397	37556528	3008	37559536	2811987968	2849547504	1984	2811992960	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
398	37573568	3168	37576736	2811974864	2849551600	46464	2812024496	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
399	37578480	3008	37581488	2812019232	2849600720	3413024	2815435264	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
400	37581936	2864	37584800	2815430992	2853015792	1984	2815435840	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
401	37585360	2768	37588128	2815431760	2853019888	2757152	2818191680	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
402	37617264	4032	37621296	2818158240	2855779536	1856	2818164128	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
403	37659664	3504	37663168	2818120464	2855783632	33791744	2851915712	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
404	37691024	2912	37693936	2851882784	2889576720	141536	2852027232	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
405	37695328	2832	37698160	2852022912	2889721072	12811136	2864836880	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
406	37698544	2608	37701152	2864833200	2902534352	2208	2864838016	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
407	37701600	2752	37704352	2864834128	2902538480	10362432	2875199312	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
408	37716720	2576	37719296	2875184112	2912903408	1952	2875188640	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
409	37724768	2944	37727712	2875179760	2912907472	864	2875183568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
410	37749280	2768	37752048	2875157568	2912909616	15744352	2890904688	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
411	37775392	3920	37779312	2890877312	2928656624	5465664	2896346896	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
412	37791264	3024	37794288	2896330464	2934124752	1984	2896335472	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
413	37799824	3520	37803344	2896325568	2934128912	1408	2896330496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
414	37807136	3280	37810416	2896322560	2934132976	3680	2896329520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
415	37814592	3632	37818224	2896320896	2934139120	3488	2896328016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
416	37822464	3120	37825584	2896319552	2934145136	1344	2896324016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
417	37834608	3552	37838160	2896309568	2934147728	4608	2896317728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
418	37842928	3200	37846128	2896308224	2934154352	1120	2896312544	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
419	37850432	3408	37853840	2896304736	2934158576	1632	2896309776	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
420	37858000	3136	37861136	2896301536	2934162672	1472	2896306144	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
421	37866208	3088	37869296	2896297472	2934166768	1280	2896301840	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
422	37929552	2752	37932304	2896238560	2934170864	25920	2896267232	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
423	37955648	3056	37958704	2896240832	2934199536	3495488	2899739376	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
424	37973808	2992	37976800	2899719696	2937696496	2528	2899725216	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
425	37980720	2896	37983616	2899716944	2937700560	1920	2899721760	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
426	37998832	2784	38001616	2899703040	2937704656	48864	2899754688	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
427	38003584	2944	38006528	2899748336	2937754864	3453632	2903204912	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
428	38006944	2448	38009392	2903201440	2941210832	3808	2903207696	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
429	38009872	2736	38012608	2903204400	2941217008	2435392	2905642528	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
430	38055008	2928	38057936	2905596224	2943654160	1888	2905601040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
431	38108384	2928	38111312	2905546912	2943658224	24871360	2930421200	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
432	38195520	3904	38199424	2930332752	2968532176	12473728	2942810384	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
433	38223248	4864	38228112	2942781504	2981009616	2342880	2945129248	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
434	38242272	2864	38245136	2945109472	2983354608	3711520	2948823856	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
435	38247584	2784	38250368	2948817264	2987067632	3730304	2952550352	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
436	38257152	2576	38259728	2952540384	2990800112	5451680	2957994640	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
437	38267296	2592	38269888	2957983504	2996253392	5726816	2963712912	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
438	38277792	2768	38280560	2963702656	3001983216	240209920	3203915344	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
439	38309616	3088	38312704	3203883536	3242196240	2180576	3206067200	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
440	38313152	2448	38315600	3206063776	3244379376	1984	3206068208	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
441	38322912	3104	38326016	3206057488	3244383504	2752	3206063344	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
442	38330832	2832	38333664	3206053872	3244387536	1504	3206058208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
443	38351760	2896	38354656	3206037008	3244391664	45632	3206085536	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
444	38367104	3072	38370176	3206068560	3244438736	2899360	3208970992	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
445	38396480	3024	38399504	3208941280	3247340784	2016	3208946320	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
446	38437408	2688	38440096	3208904752	3247344848	8260608	3217168048	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
447	38475872	2560	38478432	3217130128	3255608560	5606272	3222738960	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
448	38489328	2960	38492288	3222724752	3261217040	3532864	3226260576	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
449	38505936	3264	38509200	3226242624	3264751824	2528	3226248416	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
450	38513008	2928	38515936	3226239984	3264755920	1984	3226244896	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
451	38530208	3168	38533376	3226226672	3264760048	46304	3226276144	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
452	38535184	3664	38538848	3226269296	3264808144	3517760	3229790720	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
453	38539344	3440	38542784	3229784880	3268327664	2144	3229790464	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
454	38543360	2480	38545840	3229785920	3268331760	2417696	3232206096	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
455	38574352	2848	38577200	3232174272	3270751472	1888	3232179008	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
456	38613264	2800	38616064	3232139504	3270755568	34255616	3266397920	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
457	38645744	3136	38648880	3266364672	3305013552	141536	3266509344	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
458	38650320	3040	38653360	3266504512	3305157872	12819968	3279327520	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
459	38653936	2960	38656896	3279322448	3317979344	2464	3279327872	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
460	38657648	3616	38661264	3279322208	3317983472	10346848	3289672672	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
461	38675312	3488	38678800	3289654240	3328333040	2144	3289659872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
462	38684480	4000	38688480	3289648656	3328337136	864	3289653520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
463	38711056	3552	38714608	3289624672	3328339280	15735488	3305363712	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
464	38737776	2624	38740400	3305336640	3344077040	5454016	3310793280	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
465	38752544	3856	38756400	3310776512	3349532912	2016	3310782384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
466	38764400	3424	38767824	3310769184	3349537008	1408	3310774016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
467	38771568	3488	38775056	3310766048	3349541104	1312	3310770848	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
468	38778976	3232	38782208	3310762992	3349545200	3456	3310769680	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
469	38786480	3280	38789760	3310761584	3349551344	1312	3310766176	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
470	38798528	3984	38802512	3310753056	3349555568	4224	3310761264	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
471	38807280	2832	38810112	3310751408	3349561520	1088	3310755328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
472	38814272	2928	38817200	3310748480	3349565680	1632	3310753040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
473	38821696	3744	38825440	3310744304	3349569744	1504	3310749552	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
474	38830464	3072	38833536	3310740336	3349573872	1280	3310744688	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
475	38913568	3008	38916576	3310661360	3349577936	26720	3310691088	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
476	38942128	3168	38945296	3310661344	3349606640	3496640	3314161152	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
477	38965424	3504	38968928	3314135824	3353104752	2400	3314141728	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
478	38972752	2912	38975664	3314133056	3353108720	2112	3314138080	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
479	38991056	2992	38994048	3314118768	3353112816	47680	3314169440	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
480	38996160	3120	38999280	3314163680	3353162960	3506560	3317673360	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
481	38999728	2464	39002192	3317670016	3356672208	3680	3317676160	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
482	39002720	2512	39005232	3317673120	3356678352	2438752	3320114384	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
483	39040000	2816	39042816	3320076784	3359119600	1888	3320081488	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
484	39094368	2976	39097344	3320026320	3359123664	24909184	3344938480	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
485	39177216	3904	39181120	3344854448	3384035568	12198048	3357056400	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
486	39203856	3760	39207616	3357029168	3396236784	2509216	3359542144	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
487	39221088	3152	39224240	3359523232	3398747472	3777248	3363303632	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
488	39226560	2752	39229312	3363296688	3402526000	3770368	3367069808	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
489	39236096	2992	39239088	3367060256	3406299344	5082752	3372146000	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
490	39246496	2816	39249312	3372135248	3411384560	5710464	3377848528	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
491	39256992	3488	39260480	3377837424	3417097904	240101408	3617942320	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
492	39290672	3024	39293696	3617907216	3657200912	2166688	3620076928	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
493	39294160	2768	39296928	3620073808	3659370736	1920	3620078496	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
494	39303136	3152	39306288	3620068544	3659374832	2528	3620074224	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
495	39310944	3008	39313952	3620064944	3659378896	1536	3620069488	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
496	39330736	2976	39333712	3620049184	3659382896	45856	3620098016	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
497	39345776	3440	39349216	3620081904	3659431120	2762624	3622847968	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
498	39375168	5648	39380816	3622815104	3662195920	1920	3622822672	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
499	39418000	2976	39420976	3622779040	3662200016	8099264	3630881280	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
500	39456784	2592	39459376	3630841600	3670300976	5523968	3636368160	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
501	39469776	3200	39472976	3636353472	3675826448	3822336	3640179008	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
502	39486640	3344	39489984	3640160080	3679650064	2528	3640165952	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
503	39493872	3344	39497216	3640156912	3679654128	2080	3640162336	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
504	39511136	3280	39514416	3640143808	3679658224	46624	3640193712	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
505	39516128	3728	39519856	3640187488	3679707344	3806144	3643997360	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
506	39520288	3552	39523840	3643990928	3683514768	1920	3643996400	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
507	39524400	4752	39529152	3643989552	3683518704	2444096	3646438400	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
508	39558256	2864	39561120	3646403920	3685965040	1920	3646408704	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
509	39597840	3616	39601456	3646367680	3685969136	34983616	3681354912	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
510	39630896	3184	39634080	3681320080	3720954160	139744	3681463008	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
511	39635648	3440	39639088	3681456320	3721095408	12296512	3693756272	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
512	39639504	2592	39642096	3693752544	3733394640	9696	3693764832	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
513	39642576	2624	39645200	3693761760	3733406960	10774656	3704539040	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
514	39667072	3872	39670944	3704513616	3744184560	7232	3704524720	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
515	39676736	3872	39680608	3704514320	3744194928	3328	3704521520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
516	39703200	2864	39706064	3704494912	3744200976	15784672	3720282448	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
517	39729776	2608	39732384	3720254608	3759986992	5478976	3725736192	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
518	39744400	3600	39748000	3725720400	3765468400	1952	3725725952	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
519	39753696	3280	39756976	3725715520	3765472496	1440	3725720240	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
520	39760768	3296	39764064	3725712528	3765476592	1280	3725717104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
521	39767872	3264	39771136	3725709520	3765480656	3456	3725716240	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
522	39775184	3504	39778688	3725708112	3765486800	1312	3725712928	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
523	39787168	4032	39791200	3725699600	3765490800	4448	3725708080	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
524	39795856	3600	39799456	3725697584	3765497040	1312	3725702496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
525	39803984	3072	39807056	3725692672	3765499728	1632	3725697376	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
526	39811120	3008	39814128	3725689088	3765503216	1472	3725693568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
527	39818976	3296	39822272	3725685040	3765507312	1280	3725689616	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
528	39883376	3584	39886960	3725624448	3765511408	24768	3725652800	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
529	39910752	3136	39913888	3725624144	3765538032	3497504	3729124784	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
530	39929792	3168	39932960	3729104080	3769037040	2464	3729109712	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
531	39936864	3872	39940736	3729100400	3769041136	2016	3729106288	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
532	39955568	3680	39959248	3729085984	3769045232	48544	3729138208	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
533	39961248	3008	39964256	3729131120	3769095376	3424864	3732558992	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
534	39964912	3120	39968032	3732554736	3772522768	3552	3732561408	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
535	39968656	3120	39971776	3732557136	3772528912	2510688	3735070944	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
536	40005600	3392	40008992	3735032784	3775041776	1888	3735038064	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
537	40050128	3184	40053312	3734992560	3775045872	24845056	3759840800	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
538	40142992	3632	40146624	3759746608	3799893232	11830912	3771581152	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
539	40169056	3440	40172496	3771553984	3811726480	2388928	3773946352	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
540	40185728	3184	40188912	3773928704	3814117616	3942304	3777874192	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
541	40191168	3056	40194224	3777867808	3818062032	3950304	3781821168	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
542	40201120	3424	40204544	3781810192	3822014736	5234368	3787047984	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
543	40211856	3824	40215680	3787034736	3827250416	5660672	3792699232	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
544	40223504	3312	40226816	3792686320	3832913136	240325056	4033014688	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
545	40257104	3344	40260448	4032980400	4073240848	2157632	4035141376	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
546	40260928	2688	40263616	4035136784	4075400400	1952	4035141424	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
547	40270096	3296	40273392	4035131104	4075404496	2560	4035136960	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
548	40278304	3280	40281584	4035127040	4075408624	1504	4035131824	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
549	40298560	2992	40301552	4035111232	4075412784	50400	4035164624	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
550	40313952	3136	40317088	4035148848	4075465936	2760608	4037912592	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
551	40342080	3616	40345696	4037882160	4078227856	1952	4037887728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
552	40382288	2880	40385168	4037846624	4078231792	7826784	4045676288	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
553	40427632	2976	40430608	4045630720	4086061328	5463200	4051096896	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
554	40440912	3392	40444304	4051082048	4091526352	3570432	4054655872	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
555	40458800	3744	40462544	4054636640	4095099184	2336	4054642720	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
556	40466544	3232	40469776	4054633408	4095103184	2208	4054638848	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
557	40484256	3184	40487440	4054619872	4095107312	48384	4054671440	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
558	40489216	3200	40492416	4054665072	4095157488	4305856	4058974128	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
559	40492944	2896	40495840	4058969616	4099465456	1920	4058974432	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
560	40496384	3168	40499552	4058970032	4099469584	2451072	4061424272	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
561	40530736	3024	40533760	4061389328	4101923088	1856	4061394208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
562	40570688	3232	40573920	4061353200	4101927120	33911744	4095268176	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
563	40602272	4112	40606384	4095234624	4135841008	206816	4095445552	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
564	40607856	2976	40610832	4095440096	4136050928	12091680	4107534752	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
565	40611248	2736	40613984	4107530352	4148144336	2208	4107535296	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
566	40614416	2848	40617264	4107531200	4148148464	10751008	4118285056	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
567	40629648	2832	40632480	4118270032	4158902512	1920	4118274784	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
568	40637920	2672	40640592	4118266016	4158906608	3328	4118272016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
569	40663168	3264	40666432	4118246320	4158912752	15259232	4133508816	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
570	40689488	4000	40693488	4133480160	4174173648	5466400	4138950560	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
571	40706640	4048	40710688	4138931984	4179642672	1920	4138937952	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
572	40715904	3392	40719296	4138927408	4179646704	1440	4138932240	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
573	40723024	2944	40725968	4138924800	4179650768	1312	4138929056	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
574	40729808	3264	40733072	4138921824	4179654896	3072	4138928160	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
575	40737152	3280	40740432	4138920608	4179661040	1312	4138925200	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
576	40749472	3264	40752736	4138912304	4179665040	4800	4138920368	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
577	40757424	2800	40760224	4138911024	4179671248	1120	4138914944	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
578	40764736	2992	40767728	4138907584	4179675312	1632	4138912208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
579	40771824	2960	40774784	4138904688	4179679472	1568	4138909216	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
580	40779488	3328	40782816	4138900720	4179683536	1248	4138905296	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
581	40844480	3120	40847600	4138840096	4179687696	28672	4138871888	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
582	40869664	3360	40873024	4138845360	4179718384	3498816	4142347536	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
583	40890320	3984	40894304	4142324176	4183218480	2432	4142330592	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
584	40898320	3280	40901600	4142320912	4183222512	2112	4142326304	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
585	40917184	3712	40920896	4142305680	4183226576	48096	4142357488	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
586	40923136	3152	40926288	4142350464	4183276752	3518400	4145872016	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
587	40926736	2864	40929600	4145866864	4186796464	3456	4145873184	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
588	40930064	2896	40932960	4145869488	4186802448	2435904	4148308288	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
589	40967040	3232	40970272	4148270288	4189240560	1888	4148275408	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
590	41011104	2800	41013904	4148230752	4189244656	26090176	4174323728	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
591	41102608	4128	41106736	4174230464	4215337200	12096896	4186331488	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
592	41130176	3280	41133456	4186303552	4227437008	2347104	4188653936	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
593	41146768	3696	41150464	4188636400	4229786864	3730976	4192371072	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
594	41152720	3984	41156704	4192362608	4233519312	4067328	4196433920	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
595	41168160	3184	41171344	4196417344	4237588688	5428096	4201848624	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
596	41178784	2928	41181712	4201837312	4243019024	5745408	4207585648	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
597	41189504	2800	41192304	4207574400	4248766704	242399936	4449977136	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
598	41222176	3840	41226016	4449942032	4491168048	2168800	4452114672	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
599	41226608	2656	41229264	4452109632	4493338896	1920	4452114208	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
600	41235472	3184	41238656	4452104304	4493342960	2880	4452110368	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
601	41243360	2976	41246336	4452100848	4493347184	1504	4452105328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
602	41263632	5136	41268768	4452082416	4493351184	46592	4452134144	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
603	41281440	3328	41284768	4452114480	4493399248	2748320	4454866128	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
604	41311552	3232	41314784	4454834960	4496149744	1952	4454840144	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
605	41351648	3280	41354928	4454798944	4496153872	7929856	4462732080	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
606	41396240	3328	41399568	4462687008	4504086576	5458016	4468148352	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
607	41410160	3456	41413616	4468133120	4509546736	3492480	4471629056	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
608	41428480	3344	41431824	4471609920	4513041744	2400	4471615664	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
609	41435856	3264	41439120	4471606592	4513045712	1984	4471611840	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
610	41453632	3040	41456672	4471593168	4513049840	47232	4471643440	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
611	41458464	2768	41461232	4471638784	4513100016	4172256	4475813808	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
612	41461792	2768	41464560	4475809248	4517273808	1984	4475814000	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
613	41465120	3056	41468176	4475809760	4517277936	2537024	4478349840	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
614	41497568	5216	41502784	4478313648	4519816432	1888	4478320752	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
615	41538912	2960	41541872	4478278688	4519820560	33533088	4511814736	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
616	41568368	3408	41571776	4511783760	4553355536	157600	4511944768	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
617	41573200	3056	41576256	4511938992	4553515248	12645408	4524587456	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
618	41576672	2544	41579216	4524583424	4566162640	2432	4524588400	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
619	41579616	2624	41582240	4524584528	4566166768	10464160	4535051312	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
620	41597904	2864	41600768	4535031440	4576632208	2400	4535036704	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
621	41606352	3200	41609552	4535026592	4576636144	2208	4535032000	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
622	41631648	5152	41636800	4535003440	4576640240	15692288	4550700880	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
623	41660768	2880	41663648	4550670416	4592334064	5631424	4556304720	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
624	41675808	3024	41678832	4556289280	4597968112	2560	4556294864	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
625	41684272	3296	41687568	4556284640	4597972208	1440	4556289376	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
626	41691328	2960	41694288	4556281984	4597976272	1472	4556286416	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
627	41699664	3056	41702720	4556277648	4597980368	3680	4556284384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
628	41706768	2784	41709552	4556276928	4597986480	3136	4556282848	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
629	41718288	3824	41722112	4556270576	4597992688	6432	4556280832	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
630	41727056	2976	41730032	4556270848	4598000880	1280	4556275104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
631	41734256	3424	41737680	4556267296	4598004976	7072	4556277792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
632	41741776	2768	41744544	4556268880	4598013424	11168	4556282816	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
633	41749424	3360	41752784	4556274720	4598027504	1248	4556279328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
634	41813760	3216	41816976	4556214624	4598031600	71488	4556289328	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
635	41839776	3104	41842880	4556262416	4598105296	3607424	4559872944	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
636	41857984	2992	41860976	4559853024	4601714000	2464	4559858480	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
637	41864976	3152	41868128	4559849872	4601718000	2080	4559855104	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
638	41883168	2768	41885936	4559836160	4601722096	48672	4559887600	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
639	41887840	3056	41890896	4559881408	4601772304	3869632	4563754096	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
640	41891296	3440	41894736	4563750272	4605645008	4032	4563757744	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
641	41895280	2944	41898224	4563752928	4605651152	2434688	4566190560	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
642	41940368	4848	41945216	4566142064	4608087280	1856	4566148768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
643	41986240	3168	41989408	4566102000	4608091408	25006880	4591112048	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
644	42064880	11392	42076272	4591024256	4633100528	12395104	4603430752	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
645	42098816	3264	42102080	4603396976	4645499056	2352544	4605752784	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
646	42115104	2848	42117952	4605736368	4647854320	3720192	4609459408	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
647	42120208	3216	42123424	4609454160	4651577584	3872096	4613329472	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
648	42130048	2976	42133024	4613319472	4655452496	5580512	4618902960	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
649	42140336	3216	42143552	4618890736	4661034288	5712640	4624606592	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
650	42151568	2848	42154416	4624594752	4666749168	241404896	4866002496	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
651	42184416	3072	42187488	4865968688	4908156176	2166912	4868138672	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
652	42188048	4256	42192304	4868132672	4910324976	2016	4868138944	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
653	42198960	3168	42202128	4868126944	4910329072	2912	4868133024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
654	42207552	3536	42211088	4868122176	4910333264	1536	4868127248	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
655	42228352	3328	42231680	4868105584	4910337264	48064	4868156976	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
656	42243984	3664	42247648	4868140784	4910388432	2752704	4870897152	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
657	42273360	3472	42276832	4870866160	4913142992	1920	4870871552	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
658	42313360	2880	42316240	4870830880	4913147120	7930720	4878764480	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
659	42351024	2784	42353808	4878725344	4921079152	5452960	4884181088	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
660	42365344	3120	42368464	4884166432	4926534896	3469120	4887638672	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
661	42381936	3200	42385136	4887621216	4930006352	2464	4887626880	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
662	42389376	3136	42392512	4887617808	4930010320	2016	4887622960	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
663	42407104	3024	42410128	4887604320	4930014448	46528	4887653872	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
664	42412112	2880	42414992	4887648608	4930063600	3409568	4891061056	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
665	42415504	2912	42418416	4891057120	4933475536	1952	4891061984	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
666	42418944	2592	42421536	4891058128	4933479664	2424224	4893484944	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
667	42450368	2896	42453264	4893452256	4935905520	1888	4893457040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
668	42490640	4240	42494880	4893414768	4935909648	33606624	4927025632	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
669	42523328	3088	42526416	4926992960	4969519376	139712	4927135760	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
670	42527904	2960	42530864	4927130816	4969661680	12773024	4939906800	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
671	42531376	2512	42533888	4939902192	4982436080	2176	4939906880	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
672	42534288	2912	42537200	4939902976	4982440176	10351520	4950257408	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
673	42549168	3008	42552176	4950240800	4992792976	1920	4950245728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
674	42557632	3136	42560768	4950236144	4992796912	832	4950240112	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
675	42582272	4560	42586832	4950212224	4992799056	15773088	4965989872	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
676	42610448	2688	42613136	4965961600	5008574736	5499680	4971463968	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
677	42624080	3872	42627952	4971448704	5014076656	2496	4971455072	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
678	42633792	3760	42637552	4971443200	5014080752	1440	4971448400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
679	42641376	3296	42644672	4971440176	5014084848	3712	4971447184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
680	42648848	11888	42660736	4971430256	5014090992	3584	4971445728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
681	42664816	3616	42668432	4971428576	5014097008	1344	4971433536	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
682	42678416	3136	42681552	4971419840	5014101392	4544	4971427520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
683	42686544	2944	42689488	4971417920	5014107408	1088	4971421952	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
684	42693952	3616	42697568	4971413904	5014111472	1600	4971419120	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
685	42701728	3088	42704816	4971410720	5014115536	1920	4971415728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
686	42709776	3056	42712832	4971406832	5014119664	1248	4971411136	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
687	42774272	3312	42777584	4971346144	5014123728	26144	4971375600	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
688	42800128	3248	42803376	4971349120	5014152496	3804640	4975157008	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
689	42818640	3056	42821696	4975136944	5017958640	22080	4975162080	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
690	42825472	3040	42828512	4975154672	5017983184	32192	4975189904	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
691	42843696	2976	42846672	4975171360	5018018032	47840	4975222176	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
692	42848704	2736	42851440	4975216736	5018068176	3836736	4979056208	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
693	42851872	2672	42854544	4979051712	5021906256	3744	4979058128	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
694	42855056	2752	42857808	4979054528	5021912336	2440928	4981498208	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
695	42891168	3120	42894288	4981460288	5024354576	1856	4981465264	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
696	42934592	3792	42938384	4981420256	5024358640	24941184	5006365232	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
697	43012032	3600	43015632	5006285568	5049301200	12450592	5018739760	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
698	43037520	3328	43040848	5018714496	5061755344	2347648	5021065472	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
699	43053808	2832	43056640	5021048560	5064105200	3717760	5024769152	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
700	43058816	2960	43061776	5024762688	5067824464	3725984	5028491632	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
701	43077152	2800	43079952	5028471744	5071551696	5502592	5033977136	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
702	43087168	3392	43090560	5033965168	5077055728	5676000	5039644560	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
703	43098672	3280	43101952	5039632912	5082734864	240707424	5280343616	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
704	43131232	3136	43134368	5280310128	5323444496	2169536	5282482800	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
705	43134784	2784	43137568	5282477744	5325615312	1952	5282482480	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
706	43144112	3472	43147584	5282471856	5325619440	2880	5282478208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
707	43152112	2928	43155040	5282468560	5325623600	1536	5282473024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
708	43171808	2768	43174576	5282453056	5325627632	47104	5282502928	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
709	43187280	3328	43190608	5282486144	5325676752	2767680	5285257152	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
710	43218224	3136	43221360	5285225344	5328446704	1920	5285230400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
711	43258944	4080	43263024	5285187776	5328450800	8481312	5293673168	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
712	43297584	3776	43301360	5293632288	5336933648	5526464	5299162528	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
713	43311920	3280	43315200	5299146992	5342462192	3485760	5302636032	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
714	43330368	4160	43334528	5302615408	5345949936	2400	5302621968	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
715	43338352	2992	43341344	5302612688	5345954032	1984	5302617664	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
716	43355344	2992	43358336	5302599792	5345958128	46688	5302649472	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
717	43360288	2992	43363280	5302642944	5346006224	3422112	5306068048	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
718	43363728	3616	43367344	5306063296	5349430640	1952	5306068864	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
719	43368000	3360	43371360	5306063216	5349434576	2419200	5308485776	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
720	43401296	3168	43404464	5308450880	5351855344	1888	5308455936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
721	43447696	3296	43450992	5308408416	5351859408	34284960	5342696672	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
722	43479520	2848	43482368	5342664720	5386147088	139712	5342807280	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
723	43484000	4368	43488368	5342801024	5386289392	12727072	5355532464	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
724	43488864	2992	43491856	5355526848	5399018704	1984	5355531824	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
725	43492304	2896	43495200	5355527664	5399022864	10358368	5365888928	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
726	43507552	2944	43510496	5365873200	5409383696	1920	5365878064	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
727	43516336	2816	43519152	5365868608	5409387760	832	5365872256	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
728	43540528	2896	43543424	5365846448	5409389872	15717632	5381566976	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
729	43566800	3376	43570176	5381540080	5425110256	5460384	5387003840	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
730	43581664	3664	43585328	5386986912	5430572240	1952	5386992528	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
731	43590800	5296	43596096	5386980272	5430576368	1440	5386987008	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
732	43599648	3680	43603328	5386977136	5430580464	1312	5386982128	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
733	43607376	3168	43610544	5386974016	5430584560	3392	5386980576	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
734	43614384	3504	43617888	5386972784	5430590672	1376	5386977664	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
735	43627120	3712	43630832	5386963968	5430594800	4256	5386971936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
736	43635568	3712	43639280	5386961696	5430600976	1088	5386966496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
737	43643600	3984	43647584	5386957456	5430605040	1600	5386963040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
738	43651664	3360	43655024	5386954080	5430609104	1504	5386958944	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
739	43659968	3760	43663728	5386949600	5430613328	1248	5386954608	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
740	43724128	2992	43727120	5386890176	5430617296	25312	5386918480	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
741	43749936	3456	43753392	5386890560	5430643952	3500224	5390394240	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
742	43771760	3200	43774960	5390371104	5434146064	2496	5390376800	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
743	43781168	3520	43784688	5390365504	5434150192	2112	5390371136	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
744	43799360	3232	43802592	5390351632	5434154224	49952	5390404816	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
745	43804544	3648	43808192	5390398256	5434206448	4299072	5394700976	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
746	43808624	2944	43811568	5394696704	5438508272	3648	5394703296	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
747	43812224	2912	43815136	5394699280	5438514416	2441600	5397143792	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
748	43848848	3424	43852272	5397106432	5440958704	1888	5397111744	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
749	43892688	4320	43897008	5397065792	5440962800	24888224	5421958336	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
750	43978688	3840	43982528	5421869808	5465852336	12351616	5434225264	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
751	44004944	3200	44008144	5434205792	5478213936	2420480	5436629472	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
752	44021312	3184	44024496	5436612160	5480636656	3707936	5440323280	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
753	44026976	2896	44029872	5440316704	5484346576	3735616	5444055216	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
754	44036448	2768	44039216	5444044960	5488084176	5065568	5449113296	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
755	44047552	2672	44050224	5449101760	5493151984	5827808	5454932240	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
756	44057952	3200	44061152	5454920464	5498981616	241799776	5696723440	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
757	44096032	3168	44099200	5696683696	5740782896	2166496	5698853360	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
758	44099648	2768	44102416	5698848448	5742950864	1984	5698853200	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
759	44108928	2864	44111792	5698842912	5742954704	2912	5698848688	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
760	44116624	2832	44119456	5698839472	5742958928	1536	5698843840	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
761	44136320	3392	44139712	5698823184	5742962896	46112	5698872688	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
762	44152480	3328	44155808	5698856240	5743012048	2754048	5701613616	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
763	44186144	2880	44189024	5701579664	5745768688	1920	5701584464	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
764	44226496	3056	44229552	5701543200	5745772752	7818272	5709364528	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
765	44267568	2928	44270496	5709322608	5753593104	5449504	5714775040	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
766	44281088	3328	44284416	5714760400	5759044816	3466176	5718229904	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
767	44298464	3984	44302448	5718209856	5762512304	2304	5718216144	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
768	44306352	3072	44309424	5718206784	5762516208	1952	5718211808	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
769	44323712	4112	44327824	5718192480	5762520304	47168	5718243760	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
770	44329552	2576	44332128	5718237328	5762569456	3411264	5721651168	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
771	44332576	2672	44335248	5721648192	5765983440	1952	5721652816	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
772	44335776	2576	44338352	5721649216	5765987568	2425440	5724077232	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
773	44370160	2992	44373152	5724042352	5768415504	1888	5724047232	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
774	44409296	3216	44412512	5724007056	5768419568	34226240	5758236512	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
775	44439248	2928	44442176	5758206800	5802648976	140032	5758349760	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
776	44443744	3120	44446864	5758344288	5802791152	11884448	5770231856	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
777	44447280	2704	44449984	5770226896	5814676880	1920	5770231520	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
778	44450496	2864	44453360	5770227488	5814680848	10341376	5780571728	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
779	44465360	4704	44470064	5780555200	5825025264	1984	5780561888	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
780	44475600	3792	44479392	5780549968	5825029360	2560	5780556320	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
781	44501312	2688	44504000	5780529456	5825033456	15733664	5796265808	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
782	44527440	2752	44530192	5796238208	5840768400	5464736	5801705696	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
783	44542064	3776	44545840	5801688576	5846234416	1920	5801694272	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
784	44550864	2816	44553680	5801684800	5846238480	1408	5801689024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
785	44557152	2976	44560128	5801682384	5846242512	1312	5801686672	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
786	44564160	4992	44569152	5801677456	5846246608	3072	5801685520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
787	44573088	2736	44575824	5801676928	5846252752	1344	5801681008	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
788	44584048	3360	44587408	5801669472	5846256880	4864	5801677696	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
789	44591984	2880	44594864	5801668192	5846263056	1088	5801672160	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
790	44599088	3312	44602400	5801664752	5846267152	1664	5801669728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
791	44606464	3824	44610288	5801660896	5846271184	1504	5801666224	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
792	44615008	3696	44618704	5801656608	5846275312	1280	5801661584	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
793	44679024	3312	44682336	5801597040	5846279376	25952	5801626304	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
794	44706848	2944	44709792	5801598288	5846308080	3495520	5805096752	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
795	44727696	3312	44731008	5805075024	5849806032	2368	5805080704	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
796	44734896	4432	44739328	5805070800	5849810128	2112	5805077344	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
797	44754048	3200	44757248	5805056976	5849814224	48096	5805108272	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
798	44759184	3328	44762512	5805101920	5849864432	3438592	5808543840	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
799	44762976	2896	44765872	5808540192	5853306064	3424	5808546512	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
800	44766416	4032	44770448	5808541920	5853312368	2821952	5811367904	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
801	44803920	3072	44806992	5811329472	5856136464	1856	5811334400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
802	44847104	3072	44850176	5811290352	5856140528	25366272	5836659696	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
803	44932048	4240	44936288	5836572880	5881509168	11847616	5848424736	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
804	44958528	3408	44961936	5848398304	5893360240	2496864	5850898576	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
805	44975088	2880	44977968	5850881472	5895859440	3826176	5854710528	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
806	44980352	2576	44982928	5854705344	5899688272	3904192	5858612112	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
807	44989552	3024	44992576	5858602128	5903594704	5218816	5863823968	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
808	44999552	2656	45002208	5863812880	5908815088	5683744	5869499280	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
809	45009792	3008	45012800	5869487536	5914500336	241798560	6111289104	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
810	45041408	3248	45044656	6111256928	6156301584	2171424	6113431600	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
811	45045120	2832	45047952	6113427552	6158475504	1984	6113432368	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
812	45054720	3248	45057968	6113421632	6158479600	2912	6113427792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
813	45062864	15456	45078320	6113405472	6158483792	1536	6113422464	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
814	45095600	2976	45098576	6113389216	6158487792	48224	6113440416	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
815	45111360	2992	45114352	6113424608	6158538960	2755936	6116183536	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
816	45139568	2976	45142544	6116155104	6161297648	1952	6116160032	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
817	45178976	2816	45181792	6116119920	6161301712	7908928	6124031664	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
818	45217472	3008	45220480	6123992720	6169213200	5461504	6129457232	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
819	45230960	3536	45234496	6129442704	6174677200	3478752	6132924992	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
820	45249680	3312	45252992	6132904816	6178157808	2304	6132910432	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
821	45256880	4208	45261088	6132900816	6178161904	2016	6132907040	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
822	45275440	2896	45278336	6132887664	6178166000	45856	6132936416	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
823	45280112	3344	45283456	6132930672	6178214128	3426912	6136360928	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
824	45283920	2768	45286688	6136356784	6181643472	1984	6136361536	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
825	45287168	2848	45290016	6136357584	6181647600	2425184	6138785616	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
826	45319184	3152	45322336	6138753168	6184075504	1888	6138758208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
827	45359376	3840	45363216	6138716384	6184079600	34891392	6173611616	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
828	45390320	3296	45393616	6173578816	6218972432	141280	6173723392	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
829	45395152	4240	45399392	6173717392	6219116784	11920992	6185642624	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
830	45399808	3120	45402928	6185637312	6231040240	1952	6185642384	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
831	45403504	3584	45407088	6185637248	6231044336	10766528	6196407360	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
832	45419680	3792	45423472	6196390272	6241813744	2144	6196396208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
833	45428880	2832	45431712	6196386128	6241817840	832	6196389792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
834	45453056	3504	45456560	6196363392	6241819952	16326752	6212693648	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
835	45479728	2768	45482496	6212666096	6258148592	5467264	6218136128	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
836	45493872	3264	45497136	6218120640	6263617776	1984	6218125888	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
837	45502512	2768	45505280	6218116624	6263621904	1408	6218120800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
838	45508992	2928	45511920	6218114048	6263625968	1312	6218118288	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
839	45515936	3072	45519008	6218111056	6263630064	3264	6218117392	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
840	45522928	2848	45525776	6218110432	6263636208	1440	6218114720	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
841	45533888	4496	45538384	6218101920	6263640304	4704	6218111120	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
842	45543088	2736	45545824	6218100656	6263646480	1088	6218104480	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
843	45550000	2688	45552688	6218097856	6263650544	1600	6218102144	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
844	45556608	2880	45559488	6218095152	6263654640	1472	6218099504	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
845	45564208	13216	45577424	6218081312	6263658736	1280	6218095808	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
846	45638752	3008	45641760	6218021072	6263662832	26688	6218050768	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
847	45666032	2832	45668864	6218022640	6263691504	3502624	6221528096	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
848	45690176	3312	45693488	6221502144	6267195632	2400	6221507856	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
849	45697472	3152	45700624	6221499072	6267199696	1984	6221504208	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
850	45715520	3536	45719056	6221484512	6267203568	47584	6221535632	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
851	45721232	3152	45724384	6221529616	6267254000	3522112	6225054880	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
852	45724816	2768	45727584	6225051024	6270778608	3968	6225057760	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
853	45728112	3392	45731504	6225053248	6270784752	2439168	6227495808	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
854	45767856	3760	45771616	6227455376	6273226992	1888	6227461024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
855	45813360	3168	45816528	6227414560	6273231088	25829856	6253247584	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
856	45893152	3840	45896992	6253165520	6299062512	11811904	6264981264	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
857	45918736	3424	45922160	6264955168	6310877328	2378688	6267337280	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
858	45934992	2944	45937936	6267319616	6313257552	3908832	6271231392	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
859	45940288	2832	45943120	6271224832	6317167952	3972352	6275200016	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
860	45949744	2768	45952512	6275190960	6321143472	5239040	6280432768	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
861	45959552	2736	45962288	6280422752	6326385040	5665728	6286091216	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
862	45970176	3600	45973776	6286078944	6332052720	240203808	6526286352	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
863	46001904	3280	46005184	6526254416	6572259600	2177760	6528435456	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
864	46005776	2704	46008480	6528431184	6574439664	1952	6528435840	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
865	46015264	3520	46018784	6528424944	6574443728	2976	6528431440	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
866	46023680	3216	46026896	6528421088	6574447984	1504	6528425808	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
867	46043904	3088	46046992	6528404928	6574451920	48864	6528456880	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
868	46059200	3680	46062880	6528439312	6574502192	2758368	6531201360	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
869	46096784	3168	46099952	6531162880	6577262832	1920	6531167968	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
870	46137232	2784	46140016	6531126912	6577266928	7814496	6538944192	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
871	46175328	4000	46179328	6538904848	6585084176	5459264	6544368112	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
872	46190976	3280	46194256	6544351904	6590546160	3478464	6547833648	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
873	46207680	3472	46211152	6547814752	6594025904	2400	6547820624	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
874	46215216	2944	46218160	6547811648	6594029808	2112	6547816704	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
875	46232528	2624	46235152	6547798752	6594033904	47328	6547848704	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
876	46236912	2944	46239856	6547843168	6594083024	3426688	6551272800	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
877	46240352	2880	46243232	6551269200	6597512432	1952	6551274032	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
878	46243920	2640	46246560	6551269968	6597516528	2426080	6553698688	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
879	46276864	2928	46279792	6553664672	6599944464	1856	6553669456	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
880	46316976	2912	46319888	6553628640	6599948528	33846624	6587478176	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
881	46349280	3312	46352592	6587445280	6633797872	172064	6587620656	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
882	46354064	3120	46357184	6587615760	6633972944	12456320	6600075200	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
883	46357584	3136	46360720	6600071232	6646431952	2208	6600076576	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
884	46361200	2928	46364128	6600071952	6646436080	10598688	6610673568	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
885	46376480	2832	46379312	6610657344	6657036656	1952	6610662128	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
886	46384752	3520	46388272	6610652384	6657040656	864	6610656768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
887	46410256	2736	46412992	6610629840	6657042832	15399488	6626032064	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
888	46442944	3472	46446416	6625998240	6672444656	5615488	6631617200	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
889	46459536	4016	46463552	6631597904	6678061456	2016	6631603936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
890	46469792	4016	46473808	6631591648	6678065456	1664	6631597328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
891	46477456	3120	46480576	6631589168	6678069744	1312	6631593600	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
892	46484752	3232	46487984	6631585600	6678073584	3648	6631592480	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
893	46492128	3088	46495216	6631584480	6678079696	1344	6631588912	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
894	46503712	3120	46506832	6631576992	6678083824	11104	6631591216	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
895	46511520	2768	46514288	6631581920	6678096208	1856	6631586544	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
896	46518816	3056	46521872	6631578336	6678100208	2080	6631583472	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
897	46525936	2704	46528640	6631575664	6678104304	4032	6631582400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
898	46533408	3984	46537392	6631573056	6678110448	1504	6631578544	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
899	46601728	3792	46605520	6631509024	6678114544	62208	6631575024	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
900	46628592	3440	46632032	6631546960	6678178992	3617312	6635167712	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
901	46648448	3280	46651728	6635147200	6681798928	2560	6635153040	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
902	46655872	2816	46658688	6635144336	6681803024	2080	6635149232	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
903	46673568	3632	46677200	6635129856	6681807056	48160	6635181648	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
904	46679248	2912	46682160	6635175104	6681857264	3838528	6639016544	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
905	46683024	3328	46686352	6639011936	6685698288	4000	6639019264	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
906	46686960	3136	46690096	6639014336	6685704432	2440192	6641457664	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
907	46723408	4752	46728160	6641418512	6688146672	1920	6641425184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
908	46768544	3616	46772160	6641378608	6688150768	26015712	6667397936	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
909	46856448	3904	46860352	6667309232	6714169584	12030144	6679343280	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
910	46882336	3312	46885648	6679318336	6726203984	2354880	6681676528	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
911	46898928	3200	46902128	6681658752	6728560880	3711456	6685373408	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
912	46904528	4048	46908576	6685366352	6732274928	3963872	6689334272	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
913	46915360	3296	46918656	6689323216	6736241872	5437760	6694764272	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
914	46925888	3024	46928912	6694752448	6741681360	5725024	6700480496	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
915	46936784	3664	46940448	6700468176	6747408624	239633120	6940104960	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
916	46968976	3360	46972336	6940071776	6987044112	2163136	6942238272	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
917	46972944	2816	46975760	6942233024	6989208784	1952	6942237792	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
918	46981776	4736	46986512	6942226432	6989212944	2688	6942233856	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
919	46991440	2864	46994304	6942222736	6989217040	1504	6942227104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
920	47011248	3120	47014368	6942206736	6989221104	46848	6942256704	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
921	47026912	3472	47030384	6942239968	6989270352	2756896	6945000336	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
922	47057792	2976	47060768	6944968112	6992028880	1952	6944973040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
923	47106496	3168	47109664	6944923376	6992033040	7841664	6952768208	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
924	47147552	2928	47150480	6952725568	6999876048	5460608	6958189104	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
925	47160704	3664	47164368	6958174496	7005338864	3489408	6961667568	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
926	47178480	3328	47181808	6961648896	7008830704	2496	6961654720	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
927	47186000	3536	47189536	6961645232	7008834768	1984	6961650752	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
928	47211248	3024	47214272	6961624592	7008838864	46144	6961673760	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
929	47216032	3008	47219040	6961667952	7008886992	3410336	6965081296	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
930	47219488	2800	47222288	6965077696	7012299984	1952	6965082448	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
931	47222912	2784	47225696	6965078416	7012304112	2667616	6967748816	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
932	47256576	3376	47259952	6967714496	7014974448	3712	6967721584	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
933	47298000	4288	47302288	6967678560	7014980848	34079488	7001762336	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
934	47334512	3328	47337840	7001724864	7049062704	139456	7001867648	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
935	47339248	3456	47342704	7001861248	7049203952	12768992	7014633696	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
936	47343248	3936	47347184	7014628064	7061975248	1984	7014633984	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
937	47347664	4256	47351920	7014627456	7061979376	10367616	7024999328	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
938	47365296	3888	47369184	7024979312	7072348496	1920	7024985120	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
939	47374768	4288	47379056	7024973440	7072352496	864	7024978592	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
940	47402384	2704	47405088	7024949552	7072354640	15705376	7040657632	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
941	47428528	2752	47431280	7040630400	7088061680	5511744	7046144896	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
942	47443024	3280	47446304	7046128656	7093574960	2272	7046134208	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
943	47451792	3952	47455744	7046123248	7093578992	1408	7046128608	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
944	47459664	3392	47463056	7046120032	7093583088	3680	7046127104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
945	47467568	4512	47472080	7046117120	7093589200	3616	7046125248	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
946	47476176	2976	47479152	7046116096	7093595248	1408	7046120480	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
947	47487248	3360	47490608	7046108864	7093599472	4448	7046116672	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
948	47495664	2592	47498256	7046107360	7093605616	1088	7046111040	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
949	47502976	2848	47505824	7046103920	7093609744	1664	7046108432	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
950	47509968	3104	47513072	7046100736	7093613808	1472	7046105312	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
951	47517648	3184	47520832	7046097104	7093617936	1280	7046101568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
952	47582576	4016	47586592	7046035408	7093622000	26848	7046066272	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
953	47610544	3280	47613824	7046036816	7093650640	3741248	7049781344	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
954	47632480	3600	47636080	7049758336	7097394416	13184	7049775120	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
955	47640096	2944	47643040	7049766096	7097409136	13472	7049782512	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
956	47657824	3104	47660928	7049764208	7097425136	97696	7049865008	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
957	47662976	3776	47666752	7049858736	7097525488	3970560	7053833072	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
958	47667184	3152	47670336	7053828240	7101498576	3552	7053834944	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
959	47670816	2976	47673792	7053830960	7101504752	2437920	7056271856	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
960	47707648	3632	47711280	7056233664	7103944944	1888	7056239184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
961	47752112	5072	47757184	7056191856	7103949040	24860704	7081057632	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
962	47833824	3712	47837536	7080974192	7128811728	12400832	7093378736	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
963	47859888	3712	47863600	7093351776	7141215376	2338976	7095694464	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
964	47877008	2992	47880000	7095677328	7143557328	3707360	7099387680	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
965	47882272	3328	47885600	7099381680	7147267280	3739648	7103124656	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
966	47892112	3472	47895584	7103113392	7151008976	5512160	7108629024	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
967	47902800	4768	47907568	7108614848	7156522416	5684192	7114303808	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
968	47916352	4000	47920352	7114289200	7162209552	242227488	7356520688	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
969	47955696	3872	47959568	7356479328	7404438896	2180384	7358663584	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
970	47960080	2544	47962624	7358658256	7406620880	1952	7358662752	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
971	47969216	3056	47972272	7358652704	7406624976	2560	7358658320	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
972	47976928	2752	47979680	7358649392	7406629072	1536	7358653680	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
973	47996096	3248	47999344	7358633856	7406633200	45024	7358682128	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
974	48011680	3824	48015504	7358664800	7406680304	2751680	7361420304	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
975	48040928	3072	48044000	7361389840	7409433840	1856	7361394768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
976	48088768	3936	48092704	7361345232	7409437936	7852832	7369202000	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
977	48127616	2480	48130096	7369161984	7417292080	5460768	7374625232	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
978	48140272	3568	48143840	7374611216	7422755056	3467584	7378082368	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
979	48157952	3584	48161536	7378063856	7426225392	2368	7378069808	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
980	48165648	4000	48169648	7378059840	7426229488	2112	7378065952	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
981	48183984	3280	48187264	7378046352	7426233616	45408	7378095040	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
982	48189056	2848	48191904	7378089808	7426281712	3408320	7381500976	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
983	48192384	2816	48195200	7381496400	7429691600	1952	7381501168	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
984	48195728	2512	48198240	7381497488	7429695728	2422752	7383922752	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
985	48227520	3168	48230688	7383890896	7432121584	1856	7383895920	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
986	48267808	2784	48270592	7383855056	7432125648	34878368	7418736208	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
987	48298304	2928	48301232	7418703840	7467005072	142560	7418849328	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
988	48302704	3760	48306464	7418844112	7467150576	11880896	7430728768	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
989	48307056	2720	48309776	7430723296	7479033072	1952	7430727968	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
990	48310240	3904	48314144	7430723024	7479037168	10359680	7441086608	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
991	48326832	2960	48329792	7441069296	7489399088	6624	7441078880	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
992	48335312	3488	48338800	7441068416	7489407216	1568	7441073472	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
993	48361296	2944	48364240	7441047072	7489411312	15720800	7456770816	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
994	48387760	2720	48390480	7456744384	7505134864	5521792	7462268896	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
995	48405280	3264	48408544	7462250768	7510659312	3904	7462257936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
996	48414048	3120	48417168	7462248256	7510665424	1440	7462252816	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
997	48420864	2912	48423776	7462245776	7510669552	3008	7462251696	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
998	48427840	3168	48431008	7462244688	7510675696	3520	7462251376	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
999	48435056	2992	48438048	7462243728	7510681776	2048	7462248768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1000	48447040	5072	48452112	7462233824	7510685936	6688	7462245584	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1001	48456896	3600	48460496	7462233632	7510694128	2848	7462240080	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1002	48464656	4368	48469024	7462229264	7510698288	1600	7462235232	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1003	48473168	3184	48476352	7462225968	7510702320	2304	7462231456	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1004	48481088	3184	48484272	7462222144	7510706416	2240	7462227568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1005	48545280	2880	48548160	7462162352	7510710512	29504	7462194736	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1006	48573248	2992	48576240	7462165056	7510741296	3520864	7465688912	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1007	48592848	2864	48595712	7465669392	7514265104	5600	7465677856	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1008	48599600	2864	48602464	7465669520	7514271984	1952	7465674336	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1009	48617440	3008	48620448	7465655504	7514275952	52032	7465710544	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1010	48622416	3472	48625888	7465704528	7514330416	4260224	7469968224	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1011	48626368	2960	48629328	7469962880	7518592208	1984	7469967824	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1012	48629824	3136	48632960	7469963376	7518596336	2447712	7472414224	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1013	48670864	3104	48673968	7472372832	7521046800	1856	7472377792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1014	48723920	3072	48726992	7472323616	7521050608	25226208	7497552896	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1015	48804096	3680	48807776	7497470320	7546278096	12453792	7509927792	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1016	48831328	3248	48834576	7509900896	7558735472	2351712	7512255856	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1017	48848192	3264	48851456	7512238832	7561090288	3714304	7515956400	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1018	48853744	2640	48856384	7515950992	7564807376	3786336	7519739968	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1019	48863008	2768	48865776	7519730400	7568596176	5154144	7524887312	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1020	48873008	2720	48875728	7524877344	7573753072	5813440	7530693504	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1021	48883664	3088	48886752	7530681616	7579568368	240922976	7771607680	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1022	48918176	3008	48921184	7771571888	7820493072	2156896	7773731792	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1023	48921664	2784	48924448	7773727216	7822651664	1952	7773731952	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1024	48931216	3200	48934416	7773721312	7822655728	3360	7773727872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1025	48939184	3024	48942208	7773719632	7822661840	1568	7773724224	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1026	48958992	2896	48961888	7773703984	7822665872	51264	7773758144	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1027	48974352	3440	48977792	7773741392	7822719184	2771712	7776516544	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1028	49003200	2976	49006176	7776485968	7825492144	1952	7776490896	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1029	49042816	4128	49046944	7776449360	7825496304	7849312	7784302800	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1030	49090640	3392	49094032	7784253312	7833347344	5451136	7789707840	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1031	49104256	4160	49108416	7789691824	7838800240	3479136	7793175120	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1032	49122672	3456	49126128	7793154528	7842280656	2368	7793160352	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1033	49130368	3104	49133472	7793151280	7842284752	2144	7793156528	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1034	49148048	2976	49151024	7793137856	7842288880	47424	7793188256	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1035	49152784	3008	49155792	7793183264	7842339056	3417600	7796603872	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1036	49156272	3008	49159280	7796598880	7845758160	2016	7796603904	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1037	49159808	3152	49162960	7796599328	7845762288	2423840	7799026320	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1038	49193792	2928	49196720	7798992448	7848189168	1888	7798997264	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1039	49235024	3184	49238208	7798954800	7848193008	34270688	7833228672	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1040	49265936	2944	49268880	7833196672	7882465552	140768	7833340384	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1041	49270336	3232	49273568	7833334288	7882607856	11865984	7845203504	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1042	49274224	3792	49278016	7845197104	7894475120	2464	7845203360	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1043	49278608	2928	49281536	7845197552	7894479088	10348864	7855549344	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1044	49293728	4224	49297952	7855532752	7904830704	1952	7855538928	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1045	49303280	2912	49306192	7855528608	7904834800	2656	7855534176	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1046	49329184	3088	49332272	7855506624	7904838896	15733856	7871243568	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1047	49355536	2560	49358096	7871217632	7920575728	5463552	7876683744	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1048	49370032	3536	49373568	7876668272	7926041840	1920	7876673728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1049	49378896	3088	49381984	7876663920	7926045904	1440	7876668448	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1050	49385776	4128	49389904	7876660096	7926050000	1312	7876665536	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1051	49393696	2992	49396688	7876657408	7926054096	3520	7876663920	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1052	49400864	2912	49403776	7876656496	7926060272	1344	7876660752	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1053	49412160	4224	49416384	7876647984	7926064368	4992	7876657200	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1054	49421152	2960	49424112	7876646528	7926070640	1120	7876650608	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1055	49428624	2992	49431616	7876643024	7926074640	1632	7876647648	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1056	49435632	2832	49438464	7876640240	7926078704	1504	7876644576	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1057	49443168	3296	49446464	7876636336	7926082800	1216	7876640848	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1058	49516688	3232	49519920	7876566976	7926086896	25632	7876595840	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1059	49541744	2992	49544736	7876570832	7926115568	3505824	7880079648	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1060	49562112	2960	49565072	7880058720	7929623792	2432	7880064112	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1061	49569008	2848	49571856	7880056032	7929627888	1920	7880060800	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1062	49587456	3328	49590784	7880041200	7929631984	47648	7880092176	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1063	49592704	11806400	61399104	7868282032	7929681136	3448928	7883537360	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1064	61399680	11859920	73259600	7859873440	7933133040	4032	7871737392	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1065	73260240	3024	73263264	7859875984	7933139248	2809952	7862688960	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1066	73307664	2350480	75658144	7860293968	7935952112	1856	7862646304	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1067	75703072	3960864	79663936	7856292336	7935956272	25293632	7885546832	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1068	79746944	3917648	83664592	7877587488	7961252080	11798880	7893304016	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1069	88963632	5733888	94697520	7878357152	7973054672	2382432	7886473472	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1070	94711456	240306272	335017728	7640421872	7975439600	3919904	7884648048	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1071	335024416	2168400	337192816	7642169728	7979362544	3922944	7648261072	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1072	337203344	3712	337207056	7646080448	7983287504	5215584	7651299744	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1073	337215072	2800	337217872	7651287936	7988505808	5690784	7656981520	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1074	337230096	3168	337233264	7656967392	7994200656	241148000	7898118560	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1075	337294976	6464	337301440	7898049872	8235351312	2184512	7900240848	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1076	337301936	2723280	340025216	7897513296	8237538512	1984	7900238560	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1077	340035552	4928	340040480	7897502160	8237542640	3040	7897510128	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1078	340045936	7686896	347732832	7889815920	8237548752	1568	7897504384	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1079	347761408	5435120	353196528	7884356256	8237552784	52064	7889843440	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1080	353214192	3701616	356915808	7880690320	8237606128	2756640	7887148576	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1081	356948160	4160	356952320	7883413488	8240365808	1888	7883419536	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1082	356996944	3296	357000240	7883369664	8240369904	7952640	7891325600	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1083	357046496	5888	357052384	7891271984	8248324368	5449216	7896727088	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1084	357064880	4020800	361085680	7892689408	8253775088	3472800	7900183008	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1085	361106736	3728	361110464	7896139056	8257249520	2400	7896145184	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1086	361114448	2413456	363527904	7893725712	8257253616	1984	7896141152	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1087	363545184	5984	363551168	7893706544	8257257712	45696	7893758224	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1088	363553104	34544576	398097680	7859208160	8257305840	3431232	7897183968	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1089	398098240	141808	398240048	7862499232	8260739280	1952	7862642992	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1090	398240656	11924096	410164752	7850578720	8260743472	2425536	7864928352	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1091	410204272	4128	410208400	7852962912	8263171312	1888	7852968928	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1092	410248736	10679792	420928528	7842246880	8263175408	34385728	7887312400	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1093	420958736	4240	420962976	7876599472	8297562448	140160	7876743872	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1094	420964432	4400	420968832	7876736880	8297705712	11916032	7888657312	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1095	420969408	16172368	437141776	7872482272	8309624048	2432	7888657072	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1096	437142592	5455488	442598080	7867030064	8309628144	10800032	7883285584	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1097	442612544	3056	442615600	7877814720	8320430320	1920	7877819696	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1098	442622224	2896	442625120	7877809296	8320434416	832	7877813024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1099	442652112	3360	442655472	7877781152	8320436624	16112576	7893897088	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1100	442684560	5488	442690048	7893861104	8336551152	5465952	7899332544	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1101	442708112	3952	442712064	7899307248	8342019312	1952	7899313152	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1102	442718272	3408	442721680	7899301760	8342023440	1440	7899306608	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1103	442728144	5200	442733344	7899294192	8342027536	1312	7899300704	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1104	442738480	6032	442744512	7899287088	8342031600	3680	7899296800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1105	442749424	4256	442753680	7899284096	8342037776	1312	7899289664	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1106	442768256	7120	442775376	7899266432	8342041808	4896	7899278448	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1107	442784720	4832	442789552	7899258432	8342047984	1088	7899264352	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1108	442793920	3362736	446156656	7895895424	8342052080	1632	7899259792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1109	446161632	3696	446165328	7895890848	8342056176	1472	7895896016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1110	446172096	3840	446175936	7895884336	8342060272	1248	7895889424	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1111	446247728	3776	446251504	7895812864	8342064368	26112	7895842752	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1112	446277584	3390880	449668464	7892424576	8342093040	3493632	7899309088	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1113	449690320	3728	449694048	7895893904	8345587952	2560	7895900192	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1114	449698016	2414112	452112128	7893479888	8345592016	1984	7895895984	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1115	452128752	4144	452132896	7893463248	8345596144	47840	7893515232	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1116	452135120	25754352	477889472	7867756848	8345646320	3444576	7896955776	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1117	477889952	11822032	489711984	7859380224	8349092208	4096	7871206352	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1118	489712736	5920	489718656	7859379568	8349098224	2448192	7861833680	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1119	489767792	2294704	492062496	7859486160	8351548656	1888	7861782752	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1120	492114336	3760928	495875264	7855677520	8351552784	25738272	7885176720	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1121	495967680	3856864	499824544	7877471088	8377295632	11891808	7893219760	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1122	505233904	5665120	510899024	7878291168	8389190192	2346944	7886303232	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1123	510913488	240436896	751350384	7640188544	8391538928	3761664	7884387104	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1124	751355552	2164752	753520304	7641782848	8395303152	4026528	7647974128	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1125	753528464	3264	753531728	7645799808	8399331536	5417472	7651220544	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1126	753539184	3008	753542192	7651209408	8404751600	5663296	7656875712	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1127	753552496	3328	753555824	7656860544	8410416368	241593728	7898457600	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1128	753597488	7360	753604848	7898406944	8652011792	2164960	7900579264	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1129	753605360	2726944	756332304	7897846240	8654178544	1920	7900575104	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1130	756341584	6816	756348400	7897834240	8654182640	2976	7897844032	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1131	756353632	7728624	764082256	7890104640	8654186896	1536	7897834800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1132	764105344	5441520	769546864	7884643968	8654190832	48416	7890133904	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1133	769561696	3448912	773010608	7881231424	8654242032	2752160	7887432496	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1134	773041040	4144	773045184	7883950384	8656995568	1888	7883956416	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1135	773088336	5424	773093760	7883905904	8656999664	7942144	7891853472	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1136	773132160	3328	773135488	7891808432	8664943920	5458208	7897269968	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1137	773146880	3889008	777035888	7893368960	8670404848	3475616	7900733584	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1138	777056048	3504	777059552	7896823824	8673883376	4320	7896831648	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1139	777063680	2707872	779771552	7894118128	8673889680	19776	7896845776	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1140	779787840	3856	779791696	7894120320	8673912016	52288	7894176464	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1141	779793584	33119872	812913456	7861053856	8673967312	4322208	7898495936	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1142	812914032	187792	813101824	7865189872	8678291696	1952	7865379616	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1143	813102480	12696720	825799200	7852496560	8678295760	2444640	7867637920	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1144	825834960	5632	825840592	7854902560	8680743152	1920	7854910112	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1145	825881168	10399248	836280416	7844466832	8680747248	34577664	7889443744	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1146	836312144	3328	836315472	7879011232	8715326704	141088	7879155648	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1147	836317184	3008	836320192	7879148880	8715469072	12382976	7891534864	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1148	836320688	15581440	851902128	7875951264	8727853392	1984	7891534688	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1149	851902672	5573968	857476640	7870380720	8727857360	10443584	7886398272	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1150	857490592	3136	857493728	7880809488	8738303216	1952	7880814576	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1151	857499872	2736	857502608	7880804704	8738307312	2592	7880810032	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1152	857525312	2944	857528256	7880783152	8738311408	15308832	7896094928	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1153	857554704	2960	857557664	7896064592	8753622256	5479104	7901546656	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1154	857569936	3104	857573040	7901529664	8759102704	1952	7901534720	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1155	857578896	2768	857581664	7901525136	8759106800	1408	7901529312	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1156	857585392	2784	857588176	7901522720	8759110896	3648	7901529152	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1157	857592800	3616	857596416	7901520624	8759117040	3072	7901527312	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1158	857600800	2976	857603776	7901519408	8759123184	1376	7901523760	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1159	857616288	10272	857626560	7901500624	8759127184	4640	7901515536	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1160	857631712	6768	857638480	7901494976	8759133456	1152	7901502896	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1161	857644512	3638560	861283072	7897854192	8759137264	1600	7901494352	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1162	861287728	3328	861291056	7897850528	8759141584	1472	7897855328	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1163	861296528	3264	861299792	7897845920	8759145712	1248	7897850432	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1164	861370064	3376	861373440	7897776336	8759149776	25216	7897804928	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1165	861398816	3758960	865157776	7894018656	8759176432	3694624	7901472240	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1166	865177456	3296	865180752	7897693344	8762874096	2400	7897699040	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1167	865184752	2416176	867600928	7895277328	8762878256	2176	7897695680	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1168	867617840	3488	867621328	7895260960	8762882288	50528	7895314976	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1169	867623760	23921488	891545248	7871389232	8762934480	3638240	7898948960	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1170	891545968	12482496	904028464	7862546336	8766574800	1952	7875030784	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1171	904029136	4048	904033184	7862545744	8766578928	2442912	7864992704	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1172	904078192	2298928	906377120	7862647248	8769024368	1920	7864948096	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1173	906426704	3665792	910092496	7858935872	8769028368	25552160	7888153824	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1174	910192640	3624192	913816832	7880765456	8794582288	12077248	7896466896	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1175	919352704	5796848	925149552	7881512544	8806662096	2352160	7889661552	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1176	925163920	239341408	1164505328	7644510240	8809015568	3721376	7887573024	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1177	1164510912	2162880	1166673792	7646064976	8812738768	4098432	7652326288	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1178	1166682640	3280	1166685920	7650153968	8816839888	5346304	7655503552	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1179	1166693600	2816	1166696416	7655492848	8822189264	5725184	7661220848	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1180	1166708128	3216	1166711344	7661205184	8827916528	241606016	7902814416	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1181	1166759440	8128	1166767568	7902757696	9069525264	2174688	7904940512	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1182	1166768208	2728592	1169496800	7902205520	9071702320	1952	7904936064	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1183	1169508624	4384	1169513008	7902193376	9071706384	2720	7902200480	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1184	1169518000	7785488	1177303488	7894406928	9071710416	1536	7902193952	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1185	1177330704	5426464	1182757168	7888957344	9071714512	47232	7894431040	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1186	1182776848	3458304	1186235152	7885528512	9071763664	2758720	7891745536	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1187	1186266928	4224	1186271152	7888253248	9074524400	1952	7888259424	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1188	1186319216	3120	1186322336	7888206160	9074528496	7921504	7896130784	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1189	1186364320	3824	1186368144	7896083200	9082451344	5461920	7901548944	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1190	1186381648	3372992	1189754640	7898160608	9087915248	3486688	7905020288	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1191	1189775840	4432	1189780272	7901624768	9091405040	2400	7901631600	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1192	1189784432	2427904	1192212336	7899196800	9091409136	2080	7901626784	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1193	1192229088	4480	1192233568	7899179632	9091413200	46048	7899230160	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1194	1192235504	34629104	1226864608	7864597744	9091462352	3874528	7903101376	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1195	1226865184	142416	1227007600	7868331648	9095339248	1952	7868476016	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1196	1227008240	12762544	1239770784	7855572560	9095343344	2624832	7870959936	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1197	1239810080	4960	1239815040	7858155888	9097970928	1856	7858162704	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1198	1239856560	10376672	1250233232	7847741792	9097975024	33697120	7891815584	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1199	1250265872	4592	1250270464	7881404432	9131674896	139840	7881548864	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1200	1250272576	4880	1250277456	7881539744	9131817200	11919712	7893464336	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1201	1250278192	15756848	1266035040	7877704560	9143739600	1952	7893463360	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1202	1266036336	5823472	1271859808	7871883952	9143743760	10381664	7888089088	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1203	1271875280	3280	1271878560	7882250384	9154128944	1952	7882255616	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1204	1271885296	2800	1271888096	7882245008	9154133104	1824	7882249632	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1205	1271915792	4544	1271920336	7882215968	9154136304	15711680	7897932192	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1206	1271948400	6048	1271954448	7897896160	9169850608	5507136	7903409344	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1207	1271974016	3888	1271977904	7903381824	9175359728	1920	7903387632	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1208	1271984176	6096	1271990272	7903373552	9175363824	1568	7903381216	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1209	1271994864	4976	1271999840	7903368080	9175367920	3168	7903376224	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1210	1272004384	6688	1272011072	7903363024	9175374096	3392	7903373104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1211	1272016304	7840	1272024144	7903356000	9175380144	1344	7903365184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1212	1272038832	8480	1272047312	7903337024	9175384336	4928	7903350432	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1213	1272054176	5904	1272060080	7903330656	9175390736	1120	7903337680	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1214	1272064336	3417328	1275481664	7899912784	9175394448	1632	7903331744	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1215	1275486768	4320	1275491088	7899907552	9175398640	1696	7899913568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1216	1275498016	3408	1275501424	7899901344	9175402768	7616	7899912368	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1217	1275572976	3696	1275576672	7899836304	9175412976	26336	7899866336	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1218	1275602496	4198528	1279801024	7895640496	9175441520	3844544	7903683568	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1219	1279823024	3424	1279826448	7899461344	9179287792	2336	7899467104	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1220	1279830320	2423072	1282253392	7897038496	9179291888	2048	7899463616	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1221	1282270336	4432	1282274768	7897021216	9179295984	48768	7897074416	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1222	1282276992	24866304	1307143296	7872203888	9179347184	3828800	7900898992	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1223	1307143776	12477136	1319620912	7863557056	9183177968	4032	7876038224	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1224	1319621712	3808	1319625520	7863558592	9183184112	2423808	7865986208	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1225	1319668752	2294880	1321963632	7863647392	9185611024	1888	7865944160	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1226	1322013136	3678496	1325691632	7859923424	9185615056	24620416	7888222336	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1227	1325793664	3633296	1329426960	7880811232	9210238192	12466528	7896911056	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1228	1334578096	5874496	1340452592	7882255168	9222707760	2350880	7890480544	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1229	1340468096	238892736	1579360832	7645700784	9225061616	3725056	7888318576	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1230	1579364448	2156080	1581520528	7647267424	9228787952	3735968	7653159472	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1231	1581528416	3232	1581531648	7650993872	9232525520	5525984	7656523088	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1232	1581539200	2416	1581541616	7656511488	9238053104	5802432	7662316336	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1233	1581551008	2944	1581553952	7662304176	9243858128	242206816	7904513936	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1234	1581587376	5152	1581592528	7904474432	9486066960	2168960	7906648544	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1235	1581592960	2744896	1584337856	7903899984	9488237840	2016	7906646896	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1236	1584346400	4112	1584350512	7903891392	9488241904	2944	7903898448	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1237	1584355856	7793680	1592149536	7896096592	9488246128	1536	7903891808	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1238	1592172256	5592352	1597764608	7890485488	9488250096	45504	7896123344	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1239	1597779344	3648176	1601427520	7886869648	9488297168	2763104	7893280928	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1240	1601457584	5504	1601463088	7889599936	9491063024	1920	7889607360	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1241	1601502528	3136	1601505664	7889561456	9491067120	7944544	7897509136	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1242	1601544480	2784	1601547264	7897467152	9499014416	5450496	7902920432	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1243	1601558800	3786304	1605345104	7899121088	9504466192	3475360	7906382752	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1244	1605362560	3504	1605366064	7902576736	9507942800	2368	7902582608	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1245	1605370032	2420304	1607790336	7900156400	9507946736	2048	7902578752	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1246	1607806608	3264	1607809872	7900140960	9507950832	47552	7900191776	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1247	1607811776	34027600	1641839376	7866160608	9507999984	3421152	7903609360	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1248	1641840048	140432	1641980480	7869443792	9511424272	1952	7869586176	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1249	1641981120	12368192	1654349312	7857079024	9511428336	2483008	7871930224	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1250	1654384384	4352	1654388736	7859523920	9513912656	2368	7859530640	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1251	1654430064	10593680	1665023744	7848893104	9513916848	34376960	7893863744	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1252	1665054544	4224	1665058768	7883237696	9548296464	139616	7883381536	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1253	1665060400	4352	1665064752	7883374016	9548438768	12796832	7896175200	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1254	1665065200	15672688	1680737888	7880499824	9561237712	2240	7896174752	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1255	1680738400	5455632	1686194032	7875047808	9561241840	10357696	7890861136	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1256	1686208144	3024	1686211168	7885390480	9571601648	1920	7885395424	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1257	1686217616	3264	1686220880	7885384864	9571605744	832	7885388960	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1258	1686244048	3008	1686247056	7885360800	9571607856	15753632	7901117440	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1259	1686273504	4832	1686278336	7901084720	9587363056	5458112	7906547664	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1260	1686296208	4416	1686300624	7906522400	9592823024	1984	7906528800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1261	1686306352	4080	1686310432	7906516784	9592827216	1408	7906522272	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1262	1686314784	4704	1686319488	7906511728	9592831216	1312	7906517744	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1263	1686323696	4992	1686328688	7906506624	9592835312	3296	7906514912	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1264	1686333600	4336	1686337936	7906503520	9592841456	1344	7906509200	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1265	1686348208	6320	1686354528	7906490928	9592845456	4448	7906501696	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1266	1686360256	5248	1686365504	7906486160	9592851664	1088	7906492496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1267	1686370368	3391328	1689761696	7903094000	9592855696	1632	7906486960	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1268	1689766336	3504	1689769840	7903090048	9592859888	1472	7903095024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1269	1689775344	5264	1689780608	7903083376	9592863984	1280	7903089920	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1270	1689847168	3584	1689850752	7903017328	9592868080	25184	7903046096	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1271	1689876800	3862112	1693738912	7899155792	9592894704	3628416	7906646320	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1272	1693766128	3248	1693769376	7902764528	9596533904	17728	7902785504	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1273	1693773536	2819920	1696593456	7899961024	9596554480	11680	7902792624	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1274	1696610752	4016	1696614768	7899954080	9596568848	69088	7900027184	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1275	1696616848	25285088	1721901936	7874737536	9596639472	4118272	7904140896	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1276	1721902592	11903024	1733805616	7866954400	9600760016	3840	7878861264	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1277	1733806272	2928	1733809200	7866957024	9600766224	2419840	7869379792	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1278	1733851440	2508640	1736360080	7866827872	9603187952	1888	7869338400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1279	1736407728	3696624	1740104352	7863087696	9603192048	24930496	7891714816	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1280	1740194480	3860432	1744054912	7884069488	9628124400	12466368	7900396288	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1281	1749176464	5699216	1754875680	7885718000	9640593680	2350944	7893768160	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1282	1754890528	240183200	1995073728	7647873104	9642946832	3729024	7891785328	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1283	1995079392	2160912	1997240304	7649437920	9646678224	3748576	7655347408	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1284	1997248592	3360	1997251952	7653176160	9650428112	5354496	7658534016	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1285	1997259936	3008	1997262944	7658521776	9655784720	5783808	7664308592	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1286	1997272656	3040	1997275696	7664295616	9661571312	241193152	7905491808	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1287	1997309504	8512	1997318016	7905448368	9902766384	2168640	7907625520	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1288	1997318640	2735392	2000054032	7904883168	9904937200	1952	7907620512	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1289	2000062368	5552	2000067920	7904873376	9904941296	2944	7904881872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1290	2000073360	7796944	2007870304	7897075216	9904945520	1568	7904873728	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1291	2007892000	5452368	2013344368	7891605088	9904949456	47232	7897104688	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1292	2013360384	3693200	2017053584	7887945056	9904998640	2753088	7894391344	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1293	2017085744	3984	2017089728	7890664496	9907754224	1888	7890670368	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1294	2017130592	3728	2017134320	7890624000	9907758320	7844192	7898471920	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1295	2017175040	4112	2017179152	7898426112	9915605264	5462816	7903893040	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1296	2017191040	3990144	2021181184	7899889168	9921070352	3466848	7907346160	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1297	2021201840	3728	2021205568	7903334064	9924539632	2304	7903340096	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1298	2021209680	2415488	2023625168	7900918560	9924543728	2048	7903336096	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1299	2023641696	4512	2023646208	7900901584	9924547792	47456	7900953552	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1300	2023648080	34033792	2057681872	7866915072	9924596944	3416192	7904365056	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1301	2057682416	142256	2057824672	7870191440	9928016112	1952	7870335648	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1302	2057825312	11888304	2069713616	7858306592	9928020208	2424832	7872619728	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1303	2069748688	5904	2069754592	7860693520	9930448112	1856	7860701280	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1304	2069796128	10665632	2080461760	7849990448	9930452208	34503808	7895159888	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1305	2080493008	6576	2080499584	7884458384	9964957968	140736	7884605696	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1306	2080501376	6864	2080508240	7884593024	9965101264	12747232	7897347120	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1307	2080508704	16200080	2096708784	7881141312	9977850096	2208	7897343600	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1308	2096709632	5469088	2102178720	7875675472	9977854192	10374624	7891519184	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1309	2102194704	3648	2102198352	7886033056	9988231408	1920	7886038624	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1310	2102204848	2896	2102207744	7886027760	9988235504	2528	7886033184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1311	2102235472	3760	2102239232	7886000368	9988239600	15701312	7901705440	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1312	2102284000	6896	2102290896	7901652768	10003943664	5469120	7907128784	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1313	2102310128	6272	2102316400	7907098496	10009414896	1920	7907106688	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1314	2102324864	4672	2102329536	7907089456	10009418992	1408	7907095536	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1315	2102336016	5248	2102341264	7907081824	10009423088	1376	7907088448	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1316	2102346512	6240	2102352752	7907074432	10009427184	3136	7907083808	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1317	2102357472	6080	2102363552	7907069776	10009433328	1344	7907077200	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1318	2102377168	8256	2102385424	7907052000	10009437424	6720	7907066976	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1319	2102391616	6864	2102398480	7907047136	10009445616	1088	7907055088	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1320	2102403520	3344960	2105748480	7903701232	10009449712	1600	7907047792	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1321	2105753712	3840	2105757552	7903696256	10009453808	1472	7903701568	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1322	2105763920	3536	2105767456	7903690448	10009457904	1312	7903695296	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1323	2105835568	3552	2105839120	7903622880	10009462000	27584	7903654016	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1324	2105865296	3458736	2109324032	7900168656	10009492688	3491904	7907119296	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1325	2109346016	3472	2109349488	7903638176	10012987664	2528	7903644176	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1326	2109353424	2412672	2111766096	7901225728	10012991824	2016	7903640416	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1327	2111782944	3744	2111786688	7901209136	10012995824	48352	7901261232	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1328	2111788880	25833712	2137622592	7875423408	10013046000	4232896	7905490016	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1329	2137623056	11820816	2149443872	7867837360	10017281232	4096	7879662272	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1330	2149444448	3232	2149447680	7867839696	10017287376	2498976	7870341904	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1331	2149496496	2304640	2151801136	7867987904	10019789040	1952	7870294496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1332	2151852384	3701232	2155553616	7864239488	10019793104	24911776	7892852496	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1333	2155687008	3934192	2159621200	7885085856	10044707056	12388768	7901408816	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1334	2165039888	5668176	2170708064	7886394224	10057102288	2363744	7894426144	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1335	2170723840	242069584	2412793424	7646675648	10059469072	3784768	7892530000	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1336	2412799152	2153008	2414952160	7648303632	10063255792	3747296	7654203936	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1337	2414960832	3536	2414964368	7652040000	10067004368	5073216	7657116752	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1338	2414972144	2704	2414974848	7657105808	10072080656	5698400	7662806912	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1339	2414986000	3376	2414989376	7662790960	10077780336	242541280	7905335616	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1340	2415037552	9376	2415046928	7905276032	10320322960	2168832	7907454240	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1341	2415048112	2720096	2417768208	7904725472	10322493680	1920	7907447488	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1342	2417778112	5248	2417783360	7904714416	10322497776	2944	7904722608	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1343	2417788816	7801120	2425589936	7896912096	10322502032	1536	7904714752	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1344	2425616288	5431392	2431047680	7891458288	10322505968	49696	7896939376	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1345	2431064576	3467776	2434532352	7888024880	10322557232	2751232	7894243888	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1346	2434563984	3520	2434567504	7890743200	10325310704	1952	7890748672	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1347	2434608336	11424	2434619760	7890695040	10325314800	7992640	7898699104	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1348	2434661104	4656	2434665760	7898644528	10333310288	5463872	7904113056	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1349	2434678064	4243856	2438921920	7899853488	10338775408	3477760	7907575104	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1350	2438941040	3680	2438944720	7903310112	10342254832	2368	7903316160	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1351	2438948816	2419296	2441368112	7900890816	10342258928	2048	7903312160	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1352	2441384912	5664	2441390576	7900872448	10342263024	46048	7900924160	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1353	2441392720	33746240	2475138960	7867173216	10342312176	3447488	7904366944	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1354	2475139568	141456	2475281024	7870481008	10345762032	1952	7870624416	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1355	2475281680	12293504	2487575184	7858190944	10345766128	2423296	7872907744	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1356	2487613072	7264	2487620336	7860571648	10348191984	1920	7860580832	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1357	2487661520	10706656	2498368176	7849827904	10348196080	34563424	7895097984	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1358	2498399600	5312	2498404912	7884357376	10382762288	140576	7884503264	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1359	2498406576	3440	2498410016	7884495632	10382905648	12654624	7897153696	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1360	2498410816	15788176	2514198992	7881363200	10395562192	2688	7897154064	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1361	2514199536	5491632	2519691168	7875875152	10395566320	10480992	7891847776	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1362	2519705616	2976	2519708592	7886341440	10406050032	1920	7886346336	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1363	2519715168	2944	2519718112	7886335984	10406054096	1088	7886340016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1364	2519741632	2944	2519744576	7886312080	10406056656	15701088	7902016112	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1365	2519773408	4144	2519777552	7901982688	10421760240	5465056	7907451888	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1366	2519790416	3824	2519794240	7907434192	10427228432	1920	7907439936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1367	2519800800	2880	2519803680	7907428816	10427232496	1408	7907433104	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1368	2519807424	2880	2519810304	7907426288	10427236592	1280	7907430448	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1369	2519814960	3104	2519818064	7907422624	10427240688	3232	7907428960	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1370	2519822432	2736	2519825168	7907421696	10427246864	1344	7907425776	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1371	2519833776	3104	2519836880	7907414048	10427250928	4832	7907421984	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1372	2519842208	19840	2519862048	7907394992	10427257040	1120	7907415952	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1373	2519866368	3589920	2523456288	7903804944	10427261232	1632	7907396496	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1374	2523460880	3008	2523463888	7903801376	10427265264	1472	7903805856	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1375	2523469232	3648	2523472880	7903796480	10427269360	1280	7903801408	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1376	2523542000	3472	2523545472	7903727984	10427273456	25920	7903757376	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1377	2523570768	3566720	2527137488	7900164640	10427302128	3508896	7907240256	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1378	2527159872	3264	2527163136	7903649008	10430812144	2464	7903654736	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1379	2527167136	2416960	2529584096	7901232400	10430816496	2016	7903651376	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1380	2529601088	3776	2529604864	7901215728	10430820592	48512	7901268016	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1381	2529607056	25387232	2554994288	7875877504	10430871792	3516544	7904781280	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1382	2554994816	12074240	2567069056	7867322224	10434391280	4384	7879400848	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1383	2567069680	3744	2567073424	7867324000	10434397424	2736736	7870064480	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1384	2567117872	2300880	2569418752	7867717872	10437136624	1888	7870020640	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1385	2569464176	3672864	2573137040	7864003680	10437140720	25314144	7892990688	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1386	2573239264	3993840	2577233104	7885223072	10462456176	12035712	7901252624	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1387	2582592704	5692160	2588284864	7886212496	10474497360	2454048	7894358704	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1388	2588300928	239518224	2827819152	7649134688	10476953840	3817408	7892470320	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1389	2827823696	2154608	2829978304	7650795056	10480773360	3902560	7656852224	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1390	2829986400	3600	2829990000	7654688928	10484678928	5091520	7659784048	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1391	2829997936	2992	2830000928	7659771344	10489772272	5706880	7665481216	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1392	2830012576	3152	2830015728	7665466368	10495482096	242566784	7908036304	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1393	2830050096	6160	2830056256	7907995088	10738051344	2171840	7910173088	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1394	2830056720	2734512	2832791232	7907435024	10740226256	1984	7910171520	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1395	2832799776	4880	2832804656	7907425728	10740230384	3104	7907433712	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1396	2832810096	7807648	2840617744	7899618752	10740236496	1536	7907427936	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1397	2840638672	5434064	2846072736	7894167792	10740240528	47072	7899648928	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1398	2846088784	3462528	2849551312	7890738464	10740289776	2751712	7896952704	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1399	2849582400	4832	2849587232	7893456080	10743043312	1888	7893462800	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1400	2849627280	3936	2849631216	7893416192	10743047408	7940800	7901360928	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1401	2849670288	2944	2849673232	7901316352	10750989584	5517184	7906836480	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1402	2849684752	3338896	2853023648	7903485264	10756508912	3821792	7910645952	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1403	2853043536	3520	2853047056	7907285440	10760332496	2496	7907291456	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1404	2853051120	2736096	2855787216	7904549504	10760336720	2048	7907287648	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1405	2855803920	3024	2855806944	7904533776	10760340720	46816	7904583616	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1406	2855808912	33775376	2889584288	7870804528	10760388816	3812448	7908392352	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1407	2889584848	143824	2889728672	7874475600	10764204272	1920	7874621344	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1408	2889729328	12812560	2902541888	7861666448	10764208336	2426976	7876905984	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1409	2902578064	3920	2902581984	7864055312	10766637296	1888	7864061120	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1410	2902622992	10288288	2912911280	7853730112	10766641392	34306016	7898324416	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1411	2912943696	3824	2912947520	7888001968	10800949488	140000	7888145792	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1412	2912949200	3328	2912952528	7888139296	10801091824	12269024	7900411648	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1413	2912953056	15712128	2928665184	7884697232	10813362416	5312	7900414672	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1414	2928665776	5466656	2934132432	7879238176	10813370608	10774496	7895479328	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1415	2934147376	2928	2934150304	7889996880	10824147184	4960	7890004768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1416	2934156816	2976	2934159792	7889993632	10824153424	7264	7890003872	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1417	2934182864	3328	2934186192	7889977376	10824163568	15771104	7905751808	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1418	2934215312	6400	2934221712	7905714272	10839935984	5466784	7911187456	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1419	2934240464	3840	2934244304	7911161120	10845405424	1920	7911166880	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1420	2934250496	3472	2934253968	7911155648	10845409616	1440	7911160560	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1421	2934257936	4272	2934262208	7911151440	10845413648	3168	7911158880	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1422	2934266864	5664	2934272528	7911147232	10845419760	3168	7911156064	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1423	2934276992	4288	2934281280	7911144528	10845425808	1312	7911150128	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1424	2934293232	5344	2934298576	7911131424	10845430000	4544	7911141312	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1425	2934304944	5024	2934309968	7911126176	10845436144	1088	7911132288	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1426	2934314352	3390096	2937704448	7907735824	10845440272	1632	7911127552	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1427	2937709184	3520	2937712704	7907731632	10845444336	1472	7907736624	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1428	2937718192	3728	2937721920	7907726480	10845448400	1312	7907731520	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1429	2937789344	3808	2937793152	7907659376	10845452528	26560	7907689744	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1430	2937818272	3400384	2941218656	7904262544	10845481200	3503456	7911166384	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1431	2941239440	3504	2941242944	7907744464	10848987408	2368	7907750336	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1432	2941246800	2415072	2943661872	7905329600	10848991472	2016	7907746688	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1433	2943678448	4080	2943682528	7905313040	10848995568	48320	7905365440	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1434	2943684752	24855152	2968539904	7880505808	10849045712	3440576	7908801536	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1435	2968540400	12476560	2981016960	7871470608	10852487568	4032	7883951200	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1436	2981017792	3344	2981021136	7871472448	10852493584	2674176	7874149968	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1437	2981071056	2291840	2983362896	7871807392	10855170288	1920	7874101152	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1438	2983408976	3667232	2987076208	7868098208	10855174416	25492928	7897258368	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1439	2987171408	3637776	2990809184	7889859728	10880668912	11806688	7905304192	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1440	2996262768	5728256	3001991024	7890487904	10892478928	2395648	7898611808	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1441	3002006256	240199456	3242205712	7652670176	10894875888	3943904	7896813536	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1442	3242210576	2176672	3244387248	7654434080	10898821328	3916544	7660527296	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1443	3244395360	3216	3244398576	7658341600	10902740176	5236320	7663581136	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1444	3244406384	3088	3244409472	7663568464	10907977936	5662944	7669234496	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1445	3244420688	3776	3244424464	7669218272	10913642736	242114368	7911336416	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1446	3244467104	5552	3244472656	7911286752	11155759408	2167616	7913459920	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1447	3244473312	2875520	3247348832	7910579504	11157928336	1920	7913456944	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1448	3247358544	4768	3247363312	7910569088	11157932400	2912	7910576768	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1449	3247368800	8248896	3255617696	7902320752	11157938448	1504	7910571152	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1450	3255642128	5584384	3261226512	7896715872	11157942384	47264	7902347520	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1451	3261242768	3517104	3264759872	7893231760	11157991632	2756704	7899505568	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1452	3264791280	4240	3264795520	7895955792	11160751312	1984	7895962016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1453	3264837248	4128	3264841376	7895914096	11160755472	7890016	7903808240	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1454	3264880768	3056	3264883824	7903763648	11168647472	5505440	7909272144	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1455	3264895408	3440192	3268335600	7905818912	11174154512	3673952	7912933056	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1456	3268354400	3568	3268357968	7909472672	11177830640	2592	7909478832	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1457	3268362096	2397056	3270759152	7907075552	11177834704	2048	7909474656	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1458	3270775536	3584	3270779120	7907062752	11177841872	55936	7907122272	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1459	3270781040	34240064	3305021104	7872878112	11177899216	4036064	7911154240	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1460	3305021632	143952	3305165584	7876772288	11181937872	1984	7876918224	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1461	3305166720	12820208	3317986928	7863955072	11181942000	2423648	7879198928	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1462	3318023456	4704	3318028160	7866339696	11184367856	1888	7866346288	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1463	3318069632	10271136	3328340768	7856031184	11184371952	34814112	7901116432	73	73	cuLaunchKernelEx	296 224 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1464	3328373232	4432	3328377664	7890810352	11219188016	140544	7890955328	73	73	cudaLaunchKernel	16576 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1465	3328379472	4208	3328383680	7890947632	11219331312	11914592	7902866432	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_swiglu_bf16_kernel(const unsigned short , float , long, long)
1466	3328384240	15700800	3344085040	7887163552	11231248592	1952	7902866304	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1467	3344085840	5454672	3349540512	7881712208	11231252720	10789568	7897956448	73	73	cudaLaunchKernel	37888 1 1	256 1 1	<unnamed>::quantize_nvfp4_swiglu_bf16_kernel(const unsigned short , const float , unsigned char , unsigned char , long, long, long, long, long)
1468	3349554960	3584	3349558544	7892485088	11242043632	2112	7892490784	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1469	3349565104	3088	3349568192	7892479504	11242047696	960	7892483552	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<float>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1470	3349591824	3424	3349595248	7892454720	11242049968	16312800	7908770944	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1471	3349623040	4640	3349627680	7908737488	11258365168	5469024	7914211152	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel
1472	3349646640	6560	3349653200	7914182272	11263835472	2144	7914190976	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1473	3349659152	3408	3349662560	7914177936	11263840496	1408	7914182752	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::AUnaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1474	3349666688	4032	3349670720	7914173872	11263844592	1280	7914179184	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::floor_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1475	3349676000	9264	3349685264	7914163456	11263848720	3296	7914176016	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1476	3349690080	4416	3349694496	7914160368	11263854864	1440	7914166224	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::<unnamed>::launch_clamp_scalar(at::TensorIteratorBase &, c10::Scalar, c10::Scalar, at::native::detail::ClampLimits)::[lambda() (instance 1)]::operator ()() const::[lambda() (instance 4)]::operator ()() const::[lambda(long) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1477	3349706848	6080	3349712928	7914146000	11263858928	4320	7914156400	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1478	3349720000	3776	3349723776	7914141296	11263865072	1120	7914146192	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)2, at::native::CUDAFunctorOnSelf_add<long>, std::array<char *, (unsigned long)2>>(int, T2, T3)
1479	3349728608	3384080	3353112688	7910756480	11263869168	1600	7914142160	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::index_elementwise_kernel<(int)128, (int)4, void at::native::gpu_index_kernel<void at::native::index_kernel_impl<at::native::OpaqueType<(int)4>>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>)::[lambda(char , const char , long) (instance 1)]>(at::TensorIteratorBase &, c10::ArrayRef<long>, c10::ArrayRef<long>, const T1 &, bool)::[lambda(int) (instance 1)]>(long, T3)
1480	3353117456	3856	3353121312	7910751952	11263873264	1504	7910757312	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::CUDAFunctor_add<float>, std::array<char *, (unsigned long)3>, (int)4, TrivialOffsetCalculator<(int)2, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)2>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1481	3353126912	3968	3353130880	7910746480	11263877360	1248	7910751696	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::<unnamed>::lerp_tensor_kernel(at::TensorIteratorBase &)::[lambda() (instance 2)]::operator ()() const::[lambda() (instance 2)]::operator ()() const::[lambda(float, float, float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1482	3353199920	3456	3353203376	7910678048	11263881424	25120	7910706624	73	73	cudaLaunchKernel	3024 1 1	2 32 1	std::enable_if<!T7, void>::type internal::gemvx::kernel<int, int, float, float, float, float, (bool)0, (bool)1, (bool)1, (bool)0, (int)6, (bool)0, cublasGemvParamsEx<int, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<const float>, cublasGemvTensorStridedBatched<float>, float>>(T13)
1483	3353229040	3450912	3356679952	7907228128	11263908080	3509536	7914188576	73	73	cudaLaunchKernel	37810 1 1	32 4 1	void at::native::<unnamed>::vectorized_layer_norm_kernel<c10::BFloat16, float, (bool)1>(int, T2, const T1 , const T1 , const T1 , T2 , T2 , T1 )
1484	3356698576	3504	3356702080	7910717264	11267419344	2432	7910723200	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1485	3356706080	2420976	3359127056	7908296416	11267423472	1984	7910719376	73	73	cudaLaunchKernel	63 1 1	128 1 1	void at::native::elementwise_kernel<(int)128, (int)2, void at::native::gpu_kernel_impl_nocast<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)]>(at::TensorIteratorBase &, const T1 &)::[lambda(int) (instance 1)]>(int, T3)
1486	3359144912	3536	3359148448	7908279120	11267427568	48000	7908330656	73	73	cudaLaunchKernel	6216 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1487	3359150528	24892992	3384043520	7883433424	11267476944	3441024	7911767440	73	73	cudaLaunchKernel	256 1 1	128 1 1	void <unnamed>::partial_absmax_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , float *, long, long)
1488	3384044016	12199824	3396243840	7874676592	11270920432	3776	7886880192	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_bf16_compat_kernel(const float , float , long, float)
1489	3396244448	4320	3396248768	7874677808	11270926576	2436416	7877118544	73	73	cudaLaunchKernel	37888 1 1	256 1 1	void <unnamed>::quantize_nvfp4_modulated_bf16_kernel<float>(const unsigned short , const T1 , const T1 , const int , const float , unsigned char , unsigned char *, long, long, long, long, long, long)
1490	3396289552	2466112	3398755664	7874609056	11273364720	1856	7877077024	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1491	3398800160	3738544	3402538704	7870830112	11273368816	25923936	7900492592	73	73	cuLaunchKernelEx	296 168 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1492	3402629616	3678928	3406308544	7892986928	11299295472	11797920	7908463776	73	73	cudaLaunchKernel	529340 1 1	128 1 1	void comfy::<unnamed>::rope_kernel<__nv_bfloat16, __nv_bfloat16, __nv_bfloat16, (bool)1, (bool)1, (bool)1, (bool)1, (bool)0>(const T1 , const T1 , const T2 , const T3 , const T3 , T1 , T1 *, long, long, long, int, int, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, long, float)
1493	3411393344	5711936	3417105280	7893990704	11311095984	2385056	7902087696	73	73	cudaLaunchKernel	56 37 1	32 4 1	void at::native::reduce_kernel<(int)128, (int)4, at::native::ReduceOp<c10::BFloat16, at::native::MeanOps<c10::BFloat16, float, float, c10::BFloat16>, unsigned int, c10::BFloat16, (int)4, (int)4>>(T3)
1494	3417120208	240089808	3657210016	7656274000	11313484016	3934176	7900297984	73	73	cudaLaunchKernel	1184 56 1	512 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)32, (unsigned int)1, (bool)0, (bool)0, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1495	3657213696	2164704	3659378400	7658042896	11317421296	3955424	7664163024	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void QuantInt8Kernel<(unsigned int)128, (unsigned int)64, (unsigned int)1, (bool)0, (bool)1, __nv_bfloat16>(T6 , T6 , signed char , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1496	3659386144	3040	3659389184	7661990352	11321379536	5190528	7667183920	73	73	cudaLaunchKernel	591 56 1	1024 1 1	void TransposePadPermuteKernel<(unsigned int)128, (unsigned int)64, (bool)1, __nv_bfloat16>(T4 , T4 , unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1497	3659396960	2752	3659399712	7667172016	11326571728	5673920	7672848688	73	73	cudaLaunchKernel	56 1 128	256 1 1	void MeanScaleKernel<(unsigned int)64, (bool)0, __nv_bfloat16>(T3 , signed char , float , float , float, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int)
1498	3659410928	3568	3659414496	7672833296	11332247792	242693856	7915530720	73	73	cudaLaunchKernel	296 56 1	32 4 1	void qk_int_sv_f8_attn_kernel<(unsigned int)128, (unsigned int)64, (unsigned int)32, (unsigned int)64, (unsigned int)128, (DataType)1, (QuantGranularity)2, (QuantGranularity)2, float, (bool)1, __nv_bfloat16, (ComputeUnit)1, (MaskMode)0, (bool)0, (bool)1, (bool)0, (bool)1>(signed char , signed char , signed char , T11 , float , float , float , float , float *, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, float)
1499	3659450432	6176	3659456608	7915487408	11574944016	2174560	7917668144	73	73	cudaLaunchKernel	256 1 1	128 1 1	<unnamed>::partial_absmax_bf16_vec_kernel(const uint4 , const unsigned short , float *, long, long)
1500	3659457056	2746608	3662203664	7914917312	11577120976	1984	7917665904	73	73	cudaLaunchKernel	1 1 1	128 1 1	<unnamed>::final_scale_warp_kernel(const float , float , long, float)
1501	3662211552	4016	3662215568	7914909536	11577125104	2752	7914916304	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::bfloat16_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>>(int, T2, T3)
1502	3662221200	8087248	3670308448	7906820720	11577129168	1536	7914909504	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::unrolled_elementwise_kernel<at::native::direct_copy_kernel_cuda(at::TensorIteratorBase &)::[lambda() (instance 3)]::operator ()() const::[lambda() (instance 7)]::operator ()() const::[lambda(float) (instance 1)], std::array<char *, (unsigned long)2>, (int)4, TrivialOffsetCalculator<(int)1, unsigned int>, TrivialOffsetCalculator<(int)1, unsigned int>, at::native::memory::LoadWithCast<(int)1>, at::native::memory::StoreWithCast<(int)1>>(int, T1, T2, T4, T5, T6, T7)
1503	3670332848	5503296	3675836144	7901297120	11577133264	45664	7906846080	73	73	cudaLaunchKernel	8288 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::FillFunctor<unsigned char>, std::array<char *, (unsigned long)1>>(int, T2, T3)
1504	3675853648	3804288	3679657936	7897522432	11577180368	2752608	7904079328	73	73	cudaLaunchKernel	529536 1 1	128 1 1	void comfy::<unnamed>::quantize_nvfp4_kernel<__nv_bfloat16, __nv_fp4x2_e2m1, __nv_fp8_e4m3, (bool)1, (bool)1>(const T1 , const float , T2 , T3 , unsigned long, unsigned long, unsigned long, unsigned long, float)
1505	3679689168	5072	3679694240	7900241712	11579935952	1920	7900248704	73	73	cudaLaunchKernel	1 1 1	128 1 1	void at::native::vectorized_elementwise_kernel<(int)4, at::native::BinaryFunctor<float, float, float, at::native::binary_internal::MulFunctor<float>>, std::array<char *, (unsigned long)3>>(int, T2, T3)
1506	3679736720	4640	3679741360	7900198688	11579940048	7985536	7908188864	73	73	cuLaunchKernelEx	296 42 1	384 1 1	cutlass3x_sm120_bstensorop_s16864gemm_block_scaled_ue4m3xe2m1_ue4m3xe2m1_f32_bf16_bf16_128x128x256_1x1x1_0_tnn_align32_o_vs16_bias_bf16_relu
1507	3679780544	2752	3679783296	7908145072	11587928368	5481344	7913629168	73	73	cuLaunchKernelEx	37810 6 1	128 1 1	_gate_add_kernel

The file is too large to be shown. View raw

940 KiB Raw Permalink Blame History

940 KiB

Raw Permalink Blame History