330 lines
6.7 KiB
JSON
330 lines
6.7 KiB
JSON
{
|
|
"device": "NVIDIA GB10",
|
|
"cutlass_dsl": "4.6.2",
|
|
"metadata": {
|
|
"width": 1344,
|
|
"height": 768,
|
|
"frames": 124,
|
|
"steps": 12,
|
|
"sampler_step": 1,
|
|
"seed": 440420,
|
|
"text_tokens": 100,
|
|
"hidden_shape": [
|
|
37810,
|
|
5376
|
|
],
|
|
"segments": [
|
|
[
|
|
0,
|
|
100,
|
|
1
|
|
],
|
|
[
|
|
100,
|
|
514,
|
|
2
|
|
],
|
|
[
|
|
514,
|
|
37810,
|
|
0
|
|
]
|
|
]
|
|
},
|
|
"block_index": 24,
|
|
"capacity": 1024,
|
|
"all_equal": true,
|
|
"projection_reference_total_ms": 19.279925028483074,
|
|
"projection_ring_total_ms": 21.812459309895832,
|
|
"results": [
|
|
{
|
|
"name": "attn_out_proj",
|
|
"mnk": [
|
|
37810,
|
|
5376,
|
|
7168
|
|
],
|
|
"capacity": 1024,
|
|
"chunk_count": 37,
|
|
"final_chunk_rows": 946,
|
|
"ring_bytes": 4128768,
|
|
"parity": {
|
|
"all_chunks_equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs_max": 0.0,
|
|
"reference_checksum": 81086272.0,
|
|
"candidate_checksum": 81086267.875
|
|
},
|
|
"timing": {
|
|
"warmup": 1,
|
|
"iterations": 3,
|
|
"ring_complete_ms": 21.812459309895832,
|
|
"reference_complete_ms": 19.279925028483074,
|
|
"ring_vs_reference": 1.1313560233077324,
|
|
"improvement_percent": -13.135602330773244
|
|
},
|
|
"chunks": [
|
|
{
|
|
"start": 0,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 1024,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 2048,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 3072,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 4096,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 5120,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 6144,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 7168,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 8192,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 9216,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 10240,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 11264,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 12288,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 13312,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 14336,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 15360,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 16384,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 17408,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 18432,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 19456,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 20480,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 21504,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 22528,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 23552,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 24576,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 25600,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 26624,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 27648,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 28672,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 29696,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 30720,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 31744,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 32768,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 33792,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 34816,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 35840,
|
|
"rows": 1024,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
},
|
|
{
|
|
"start": 36864,
|
|
"rows": 946,
|
|
"equal": true,
|
|
"max_abs": 0.0,
|
|
"mean_abs": 0.0
|
|
}
|
|
]
|
|
}
|
|
],
|
|
"projection_total_improvement_percent": -13.135602330773244
|
|
}
|