sglang

aarch64

[sgl-kernel/cpu]: exclude amx gemm source from arm build (#29286 )

2026-06-26 08:23:43 +08:00

mamba

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

model

[CPU] add fused input proj for qwen3.5 (#31171 )

2026-07-15 15:06:24 +08:00

x86_64

[cpu] Implement all gather/reduce for arm64 cpu (#12527 )

2025-12-08 19:03:04 +08:00

activation.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

bmm.cpp

[CPU] remove RECORD_FUNCTION (#23528 )

2026-04-24 09:18:29 +08:00

CMakeLists.txt

[sgl-kernel/cpu]: exclude amx gemm source from arm build (#29286 )

2026-06-26 08:23:43 +08:00

common.h

[CPU] Add GPT-OSS model optimization for CPU (#16775 )

2026-05-29 16:05:26 +08:00

conv3d.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

decode.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

extend.cpp

Support speculative decoding on CPU (#27862 )

2026-07-09 10:27:09 +08:00

flash_attn.cpp

[Diffusion] Revert CPU AMX optimizations (#30716 )

2026-07-10 09:09:38 +08:00

flash_attn.h

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

gemm_fp8.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

gemm_int4.cpp

[CPU] remove RECORD_FUNCTION (#23528 )

2026-04-24 09:18:29 +08:00

gemm_int8.cpp

[sgl-kernel/cpu]: exclude amx gemm source from arm build (#29286 )

2026-06-26 08:23:43 +08:00

gemm.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

gemm.h

[CPU] Add GPT-OSS model optimization for CPU (#16775 )

2026-05-29 16:05:26 +08:00

interface.cpp

[sgl-kernel][CPU] add kernel for shm_allgather_into_tensor and shm_reduce_scatter_tensor (#13397 )

2026-07-23 09:19:38 +08:00

kvcache.cpp

Support speculative decoding on CPU (#27862 )

2026-07-09 10:27:09 +08:00

moe_fp8.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

moe_int4.cpp

[CPU] expand the interface of shared_expert without scaling factor (#22933 )

2026-04-21 20:03:39 +08:00

moe_int8.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

moe.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

moe.h

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

norm.cpp

fix arm test_norm.py error (#30597 )

2026-07-09 14:36:45 +08:00

numa_utils.cpp

[CPU] improve numa memory binding (#19666 )

2026-03-18 22:15:50 -07:00

preprocessor.cpp

[CPU] remove RECORD_FUNCTION (#23528 )

2026-04-24 09:18:29 +08:00

qkv_proj.cpp

[CPU] improve silu performance by replacing fp32 div with rcp14 (#31304 )

2026-07-17 12:02:37 +08:00

rope.cpp

[CPU] refactor rope kernels (#31897 )

2026-07-22 09:12:02 +08:00

shm.cpp

[sgl-kernel][CPU] add kernel for shm_allgather_into_tensor and shm_reduce_scatter_tensor (#13397 )

2026-07-23 09:19:38 +08:00

shm.h

[sgl-kernel][CPU] add kernel for shm_allgather_into_tensor and shm_reduce_scatter_tensor (#13397 )

2026-07-23 09:19:38 +08:00

spec.cpp

Support speculative decoding on CPU (#27862 )

2026-07-09 10:27:09 +08:00

topk.cpp

[CPU] remove RECORD_FUNCTION (#23528 )

2026-04-24 09:18:29 +08:00

torch_extension_cpu.cpp

[sgl-kernel][CPU] add kernel for shm_allgather_into_tensor and shm_reduce_scatter_tensor (#13397 )

2026-07-23 09:19:38 +08:00

vec_pack.h

[CPU] optimize flash_attn_varlen_func (#15708 )

2026-01-29 22:07:05 -08:00

vec.h

[CPU] refactor rope kernels (#31897 )

2026-07-22 09:12:02 +08:00