[Kernel] add Flashinfer cutedsl w4a16 linear (#53014)

Signed-off-by: Siyuan Fu <siyuanf@nvidia.com>
This commit is contained in:
Lain
2026-09-01 02:49:14 +00:00
committed by GitHub
parent 907b1a7f22
commit 882ca8d696
8 changed files with 230 additions and 10 deletions
+1 -1
View File
@@ -276,7 +276,7 @@ class KernelConfig:
- "cutlass": Use CUTLASS-based kernels
- "flashinfer_cutlass": Use FlashInfer with CUTLASS kernels
- "flashinfer_cutedsl": Use FlashInfer with CuTe-DSL kernels
(BF16, NVFP4, MXFP8)
(BF16, NVFP4, MXFP8, W4A16_NVFP4)
- "flashinfer_trtllm": Use FlashInfer with TensorRT-LLM kernels
- "flashinfer_cudnn": Use FlashInfer with cuDNN kernels
- "flashinfer_b12x": Use FlashInfer b12x CuteDSL NVFP4 GEMM (SM120+)