Commits · 3ded4b215df396061d588eb632385bb94dc97b13 · change / sglang · GitLab

17 Mar, 2025 2 commits
- Support serving DeepSeek-R1-Channel-INT8 with 32 L40S. (#4418) · 75b65648
  Wenbo Yang authored Mar 17, 2025
  
  75b65648
- [ROCm] enable moe topk softmax in amd (#4448) · 9b8333d9
  yiakwy-xpu-ml-framework-team authored Mar 17, 2025
  
  9b8333d9
16 Mar, 2025 7 commits
- fix custom allreduce performance/accuracy problem (#4477) · 25e1816e
  Yi Zhang authored Mar 17, 2025
  
  25e1816e
- [Eagle] Remove the greedy branch and some redundant code (#4363) · 1b859295
  Ying Sheng authored Mar 16, 2025
```
Co-authored-by: Sehoon Kim <sehoon@x.ai>
```
  1b859295
- Revert "feat: Add FlashMLA submodule (#4449)" (#4470) · 9971dc22
  Yineng Zhang authored Mar 16, 2025
  
  9971dc22
- Release sgl-kernel v0.0.5.post2 (#4469) · 3db35c1a
  Lianmin Zheng authored Mar 16, 2025
  
  3db35c1a
- Add greedy verification kernel (#4383) · 52a34d74
  Ying Sheng authored Mar 16, 2025
  
  52a34d74
- [Fix] use `torch.cat` instead of `torch.concat` to prevent entering the... · 1a3fa75f
  JieXin Liang authored Mar 16, 2025
```
[Fix] use `torch.cat` instead of `torch.concat` to prevent entering the `Autograd` backends. (#4466)
```
  1a3fa75f
- feat: Add FlashMLA submodule (#4449) · 81f431ed
  Shi Shuai authored Mar 16, 2025
```
Co-authored-by: sleepcoo <sleepcoo@gmail.com>
```
  81f431ed
14 Mar, 2025 2 commits
- bump v0.0.5.post1 (#4437) · 862fe522
  Yineng Zhang authored Mar 14, 2025
  
  862fe522
- Add moe topk softmax templated from vllm (#4302) · 61e4433c
  Qingquan Song authored Mar 14, 2025
  
  61e4433c
13 Mar, 2025 5 commits
- Update pyproject.toml · 4ff12642
  Yineng Zhang authored Mar 13, 2025
  
  4ff12642
- bump 0.0.5 sgl-kernel (#4377) · 2a4cbad8
  Yineng Zhang authored Mar 13, 2025
  
  2a4cbad8
- fix accuracy issue (#4376) · 2937387a
  Yineng Zhang authored Mar 13, 2025
  
  2937387a
- Fix per token fp8 quant precision (#4362) · 4068e012
  Qingquan Song authored Mar 12, 2025
  
  4068e012
- feat: support ep size < 32 for sgl kernel (#4348) · 817d4370
  Shi Shuai authored Mar 13, 2025
  
  817d4370
12 Mar, 2025 6 commits
- Support Blackwell Block Scale FP8 Gemm (#4278) · 7c866711
  Elfie Guo authored Mar 12, 2025
  
  7c866711
- release 0.0.4.post3 sgl-kernel (#4331) · 6e7239f9
  Yineng Zhang authored Mar 12, 2025
  
  6e7239f9
- fix awq_dequantize (#4333) · 0a3960f2
  Yineng Zhang authored Mar 12, 2025
  
  0a3960f2
- Add awq dequantize kernel to sgl with 1x to 3x speedup (#4104) · 07f94463
  Rex authored Mar 12, 2025
  
  07f94463
- Remove vllm ops scaled fp8 quant and accelerate per token quant by 20-28% (#4215) · e0917e6b
  Stefan He authored Mar 12, 2025
```
Co-authored-by: Stefan He <bhe@linkedin.com>
```
  e0917e6b
- refine sgl_moe_align_block_size_benchmark (#4327) · 7130a7ce
  Xiaoyu Zhang authored Mar 12, 2025
  
  7130a7ce
11 Mar, 2025 4 commits
- [AMD] Fix rocm sgl-kernel missing modules error (#4311) · 690e1f23
  yigex authored Mar 12, 2025
```
Co-authored-by: yiakwy-xpu-ml-framework-team <leiwang2@amd.com>
```
  690e1f23
- bump sgl-kernel 0.0.4.post2 (#4288) · cd909455
  Yineng Zhang authored Mar 11, 2025
  
  cd909455
- update deepgemm (#4284) · bde24ab3
  Yineng Zhang authored Mar 10, 2025
  
  bde24ab3
- Uupdate cutalss dependency for its bug fix (#4277) · bf2eefc0
  Elfie Guo authored Mar 10, 2025
  
  bf2eefc0
10 Mar, 2025 9 commits
- add THIRDPARTYNOTICES for DeepGEMM (#4272) · 3dd4feae
  Yineng Zhang authored Mar 10, 2025
  
  3dd4feae
- Optimize rope in sgl kernel (#4267) · cf0ccd40
  Lianmin Zheng authored Mar 10, 2025
  
  cf0ccd40
- Release sgl-kernel v0.0.4.post1 (#4255) · 1a5023e0
  Lianmin Zheng authored Mar 10, 2025
  
  1a5023e0
- fix per_token_group_quant_fp8 illegal memory when num_groups % 16 != 0 (#4231) · 23308a90
  Xiaoyu Zhang authored Mar 10, 2025
  
  23308a90
- Simplify tests & Fix trtllm custom allreduce registration (#4252) · aa957102
  Lianmin Zheng authored Mar 10, 2025
  
  aa957102
- DeepGemm integrate to sgl-kernel (#4165) · c553e160
  laixin authored Mar 10, 2025
```
Co-authored-by: sleepcoo <sleepcoo@gmail.com>
Co-authored-by: HandH1998 <1335248067@qq.com>
Co-authored-by: shuaills <shishuaiuoe@gmail.com>
Co-authored-by: yinfan98 <1106310035@qq.com>
Co-authored-by: Yineng Zhang <me@zhyncs.com>
```
  c553e160
- Move activation.cu to sgl-kernel/elementwise (#4250) · 7c0541b3
  Lianmin Zheng authored Mar 09, 2025
  
  7c0541b3
- Minor style fix for sgl-kernel (#4243) · 730d084f
  Lianmin Zheng authored Mar 09, 2025
  
  730d084f
- Move rope and bmm into sgl-kernel (#4241) · eb06dbcb
  Lianmin Zheng authored Mar 09, 2025
  
  eb06dbcb
09 Mar, 2025 4 commits
- update sgl-kernel 3rdparty (#4228) · df84ab2a
  Yineng Zhang authored Mar 09, 2025
  
  df84ab2a
- chore: bump v0.0.4 for sgl-kernel (#4223) · 5c7dd14b
  Yineng Zhang authored Mar 08, 2025
  
  5c7dd14b
- Rename files in sgl kernel to avoid nested folder structure (#4213) · 8abf74e3
  Lianmin Zheng authored Mar 08, 2025
```
Co-authored-by: zhyncs <me@zhyncs.com>
```
  8abf74e3
- revert pr 3628 to pass test_mla ci (#4219) · 79a321af
  Xiaoyu Zhang authored Mar 09, 2025
  
  79a321af
08 Mar, 2025 1 commit
- refine quant kernel code style (#4211) · b3251e9f
  Xiaoyu Zhang authored Mar 08, 2025
  
  b3251e9f