jy-song-hub 4fc09e0df0 Fp4 MOE quant kernel optimization (#8777) ...

Co-authored-by: Rain Jiang <96632942+rainj-me@users.noreply.github.com>

2025-08-15 01:46:16 -07:00

..

gptq

[2/n]decouple quantization implementation from vLLM dependency (#8112)

2025-08-14 03:19:03 -07:00

marlin

[2/n]decouple quantization implementation from vLLM dependency (#8112)

2025-08-14 03:19:03 -07:00

awq_kernel.cu

Fix AWQ Dequant and Weight Loading of deepseek v2 (#6842)

2025-06-17 13:45:10 -07:00

bmm_fp8.cu

…

dsv3_fused_a_gemm.cu

Fix typos and unify size(s)/stride(s) API calls (#8799)

2025-08-08 00:18:08 -07:00

dsv3_router_gemm_bf16_out.cu

[Kimi K2] dsv3_router_gemm supports NUM_EXPERTS == 384 (#8013)

2025-08-01 22:01:24 +08:00

dsv3_router_gemm_entry.cu

[Kimi K2] dsv3_router_gemm supports NUM_EXPERTS == 384 (#8013)

2025-08-01 22:01:24 +08:00

dsv3_router_gemm_float_out.cu

[Kimi K2] dsv3_router_gemm supports NUM_EXPERTS == 384 (#8013)

2025-08-01 22:01:24 +08:00

fp8_blockwise_gemm_kernel.cu

[sgl-kernel] 1/N Refactor sglang cutlass 3x - gemm fp8 blockwise sm90 (#8913)

2025-08-14 10:55:54 -07:00

fp8_gemm_kernel.cu

[Perf] Tunings for SM100 FP8 CUTLASS kernel (#8818)

2025-08-13 21:59:22 -07:00

int8_gemm_kernel.cu

Fix typos and unify size(s)/stride(s) API calls (#8799)

2025-08-08 00:18:08 -07:00

math.hpp

[Perf] Tunings for SM100 FP8 CUTLASS kernel (#8818)

2025-08-13 21:59:22 -07:00

nvfp4_expert_quant.cu

Fp4 MOE quant kernel optimization (#8777)

2025-08-15 01:46:16 -07:00

nvfp4_quant_entry.cu

[1/2] Add Kernel support for Cutlass based Fused FP4 MoE (#6093)

2025-06-02 13:48:03 -07:00

nvfp4_quant_kernels.cu

…

nvfp4_scaled_mm_entry.cu

…

nvfp4_scaled_mm_kernels.cu

Fix typos and unify size(s)/stride(s) API calls (#8799)

2025-08-08 00:18:08 -07:00

per_tensor_quant_fp8.cu

[sgl-kernel performace] fix fp8 quant kernels dispatch __nv_fp8_e4m3 bug to improve performance 10%-20% (#8499)

2025-07-29 23:31:54 +08:00

per_token_group_quant_8bit.cu

[sgl-kernel performace] fix fp8 quant kernels dispatch __nv_fp8_e4m3 bug to improve performance 10%-20% (#8499)

2025-07-29 23:31:54 +08:00

per_token_quant_fp8.cu

[sgl-kernel] avoid per_token_quant_fp8.cu hardcode sm_count (#8738)

2025-08-04 12:55:57 +08:00

qserve_w4a8_per_chn_gemm.cu

…

qserve_w4a8_per_group_gemm.cu

…