4 people 5bb8d2767a [Kernel] Batch invariant NVFP4 linear using cutlass (#39912) ...

Signed-off-by: Jakub Zakrzewski <jzakrzewski@nvidia.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>

2026-05-23 09:41:12 -04:00

..

activation_nvfp4_quant_fusion_kernels.cu

[4/n] Migrate FP4/W4A8 CUTLASS kernels to torch stable ABI (#37503)

2026-03-31 10:21:13 -07:00

mxfp4_blockwise_moe_kernel.cu

[Kernel] Add MXFP4 W4A4 CUTLASS MoE kernel for SM100 (#37463)

2026-04-17 16:42:32 -07:00

mxfp4_experts_quant.cu

[Bugfix] Guard mxfp4_experts_quant bindings on ENABLE_NVFP4_SM100 (#40191)

2026-04-18 13:58:46 -07:00

nvfp4_blockwise_moe_kernel.cu

[4/n] Migrate FP4/W4A8 CUTLASS kernels to torch stable ABI (#37503)

2026-03-31 10:21:13 -07:00

nvfp4_experts_quant.cu

[Bugfix][Kernel] nvfp4 cutlass MoE: fix nvfp4 experts quant out-of-bounds read for expert counts not divisible by 4 or 16 (#40351)

2026-04-21 19:06:09 +00:00

nvfp4_quant_entry.cu

[Compile] Fix nvfp4 compile warning (#38573)

2026-04-01 18:28:57 +00:00

nvfp4_quant_kernels.cu

[Perf] Padded nvfp4 quant kernel to remove additional copy, 2.4%~5.7% e2e performance improvement (#42774)

2026-05-18 16:38:12 -07:00

nvfp4_scaled_mm_entry.cu

[4/n] Migrate FP4/W4A8 CUTLASS kernels to torch stable ABI (#37503)

2026-03-31 10:21:13 -07:00

nvfp4_scaled_mm_kernels.cu

[Kernel] Batch invariant NVFP4 linear using cutlass (#39912)

2026-05-23 09:41:12 -04:00

nvfp4_scaled_mm_sm120_kernels.cu

[Kernel] Batch invariant NVFP4 linear using cutlass (#39912)

2026-05-23 09:41:12 -04:00

nvfp4_utils.cuh

[4/n] Migrate FP4/W4A8 CUTLASS kernels to torch stable ABI (#37503)

2026-03-31 10:21:13 -07:00