Ming Yang fba8906930

[perf] Use direct copy (broadcast) instead of cat for k_nope/k_pe in MLA prefill (#29710) ...

Signed-off-by: Ming Yang <minos.future@gmail.com>

2025-12-11 08:20:45 +00:00

..

deepgemm

[Doc]: fix typos in various files (#29230)

2025-11-24 11:10:48 +00:00

bench_block_fp8_gemm.py

Remove weight_scale.T special case for SM90 Block FP8 CUTLASS kernel (#28431)

2025-11-11 11:46:04 -05:00

bench_fp8_gemm.py

…

bench_int8_gemm.py

…

bench_mxfp4_qutlass.py

…

bench_nvfp4_gemm.py

…

bench_nvfp4_qutlass.py

…

bench_per_token_quant_fp8.py

…

benchmark_2d_silu_mul_fp8_quant.py

[Performance][DP/EP] Add silu_mul_per_token_group_quant_fp8_colmajor kernel (#29470)

2025-12-03 18:04:59 +00:00

benchmark_activation.py

…

benchmark_bitblas.py

…

benchmark_cutlass_fp4_moe.py

…

benchmark_cutlass_moe_fp8.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_device_communicators.py

…

benchmark_fused_collective.py

[PERF] Allreduce fusion. Support torch native matching. Tuning of the thresholds (#24248)

2025-11-10 18:33:11 -05:00

benchmark_grouped_gemm_cutlass.py

Disable nm-testing models with issues in CI (#28206)

2025-11-06 06:19:07 -08:00

benchmark_layernorm.py

…

benchmark_lora.py

…

benchmark_machete.py

[Kernel][Quantization] add w4a8 support for marlin kernel (#24722)

2025-11-29 07:19:33 -08:00

benchmark_marlin.py

[Kernel][Quantization] add w4a8 support for marlin kernel (#24722)

2025-11-29 07:19:33 -08:00

benchmark_mla_k_concat.py

[perf] Use direct copy (broadcast) instead of cat for k_nope/k_pe in MLA prefill (#29710)

2025-12-11 08:20:45 +00:00

benchmark_moe_align_block_size.py

[Kernel][MoE] optimize moe_align_block_size (#29642)

2025-12-07 01:58:47 -08:00

benchmark_moe_permute_unpermute.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_moe.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_mrope.py

Update rope_scaling to rope_parameters in preparation for Transformers v5 (#28542)

2025-11-19 09:06:36 -08:00

benchmark_paged_attention.py

…

benchmark_per_token_group_quant.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_quant.py

…

benchmark_reshape_and_cache_flash.py

…

benchmark_reshape_and_cache.py

…

benchmark_rmsnorm.py

…

benchmark_rope.py

Fix rotary embedding benchmark script (#28323)

2025-11-10 21:57:12 -05:00

benchmark_shapes.py

Disable nm-testing models with issues in CI (#28206)

2025-11-06 06:19:07 -08:00

benchmark_silu_mul_fp8_quant.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_trtllm_decode_attention.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_trtllm_prefill_attention.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

benchmark_w8a8_block_fp8.py

Replace torch.cuda.Event with torch.Event for better hardware compatibility (#26985)

2025-11-18 11:34:36 -08:00

graph_machete_bench.py

…

requirements.txt

…

utils.py

…

weight_shapes.py

…