3 people c2eae96c56 MSCCL++ Integration (#22734) ...

Co-authored-by: Caio Rocha <caiorocha@microsof.com>
Co-authored-by: empyreus <rjsouza1995@gmail.com>

2026-06-08 21:13:13 -07:00

..

quantization

[7/n] decouple quantization impl from vllm dependency - gguf kernel (#11019)

2025-10-11 14:04:57 -07:00

testing

[kernel slimming] Clean many useless sgl-kernel deprecated kernels (#20277)

2026-03-14 16:45:54 +08:00

__init__.py

[AMD] Dsv4/pr2 compressor opt (#26208)

2026-05-25 23:54:40 -07:00

allreduce.py

MSCCL++ Integration (#22734)

2026-06-08 21:13:13 -07:00

attention.py

Remove obsolete sgl-kernel legacy paths (#21528)

2026-04-01 09:00:20 +08:00

cutlass_moe.py

[1/n]: add cutlass W4A8 moe kernel for hopper architecture (#7772)

2025-07-04 20:50:12 -07:00

debug_utils.py

Add SGLang CUDA crash API logging inspired by FlashInfer (#20910)

2026-03-22 16:39:40 +08:00

elementwise.py

[AMD] Dsv4/pr1 fix run time issue (#25898)

2026-05-23 16:04:14 -07:00

expert_specialization.py

[sgl-kernel][Feat][B200][1/N] Support MXFP8 Grouped GEMM in Blackwell (#13731)

2025-12-04 10:09:09 +08:00

flash_attn.py

perf: eliminate attention DtoD copy by passing pre-allocated output to FA (#21985)

2026-04-24 12:05:16 -07:00

flash_mla.py

Sgl flashmla (#26132)

2026-05-26 12:00:23 -07:00

gemm.py

[Kernel] Set sgl_per_token_group_quant_8bit_v2 as default choice (#22467)

2026-04-11 01:59:57 -07:00

grammar.py

fix sgl-kernel unit tests (#5666)

2025-04-23 01:18:30 -07:00

kvcacheio.py

[HiCache]Page head layout IO kernel (#11615)

2025-10-26 15:53:50 +08:00

load_utils.py

Update kernel installation instructions after shifting default cuda to 13 (#24181)

2026-05-02 16:07:20 -07:00

mamba.py

[CPU] Optimize Qwen3-next model on CPU (#12525)

2026-01-29 22:03:58 -08:00

memory.py

[kernel slimming] Clean many useless sgl-kernel deprecated kernels (#20277)

2026-03-14 16:45:54 +08:00

metal.py

[Apple Silicon] Add custom Metal RoPE kernel with fused KV cache store (#22868)

2026-05-29 15:09:33 +08:00

moe.py

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

2026-03-07 10:31:08 +08:00

musa.py

[MUSA]: Add flashinfer sampling backend (#24978)

2026-05-14 20:23:15 -07:00

sampling.py

Remove obsolete sgl-kernel legacy paths (#21528)

2026-04-01 09:00:20 +08:00

scalar_type.py

refactor: Move scalar_types.py to sgl-kernel to avoid circular import (#8720)

2025-08-07 19:22:16 -07:00

sparse_flash_attn.py

[Feat] QWen-1M context support[1/2]: Update block sparse attention backend utils kernel (#5847)

2025-04-28 11:03:17 -07:00

spatial.py

Optional extension for green context (#9231)

2025-08-15 21:33:52 +08:00

speculative.py

move all get_stream in sgl_kernel to c++ to reduce the launch overhead (#12521)

2025-11-02 13:15:05 -08:00

test_utils.py

Reland [1/2] Optimizations and refactors about quant kernel (#10312)

2025-10-11 15:59:03 +08:00

top_k.py

[AMD] Dsv4/pr1 fix run time issue (#25898)

2026-05-23 16:04:14 -07:00

utils.py

[MUSA] Use MUSA-optimized operators in piecewise CUDA graph (#23633)

2026-05-11 17:55:27 -07:00

version.py

chore: bump sgl-kernel version to 0.4.3 (#26414)

2026-05-26 12:10:16 -07:00