[JIT Kernel] Reland NVFP4 kernels to JIT (#20012)

This commit is contained in:
Mohammad Miadh Angkad
2026-03-07 10:31:08 +08:00
committed by GitHub
parent 6ffc74efd7
commit f88acf8780
27 changed files with 2596 additions and 991 deletions
-6
View File
@@ -286,11 +286,6 @@ set(SOURCES
"csrc/gemm/fp8_blockwise_gemm_kernel.cu"
"csrc/gemm/fp8_gemm_kernel.cu"
"csrc/gemm/int8_gemm_kernel.cu"
"csrc/gemm/nvfp4_expert_quant.cu"
"csrc/gemm/nvfp4_quant_entry.cu"
"csrc/gemm/nvfp4_quant_kernels.cu"
"csrc/gemm/nvfp4_scaled_mm_entry.cu"
"csrc/gemm/nvfp4_scaled_mm_kernels.cu"
"csrc/gemm/per_tensor_quant_fp8.cu"
"csrc/gemm/per_token_group_quant_8bit.cu"
"csrc/gemm/per_token_group_quant_8bit_v2.cu"
@@ -316,7 +311,6 @@ set(SOURCES
"csrc/moe/moe_sum_reduce.cu"
"csrc/moe/moe_topk_softmax_kernels.cu"
"csrc/moe/moe_topk_sigmoid_kernels.cu"
"csrc/moe/nvfp4_blockwise_moe.cu"
"csrc/moe/fp8_blockwise_moe_kernel.cu"
"csrc/moe/prepare_moe_input.cu"