Commits · 34c3f9b2d32787a5be6ab344904cb6ce418a3c8b · change / sglang

23 Jun, 2025 2 commits

kvcache io kernels and test case (#7382) · 34c3f9b2
Zhiqiang Xie authored Jun 23, 2025

34c3f9b2

feat: integrate deepgemm into EPMoE (#6821) · 506c4928

xutizhou authored Jun 23, 2025


Co-authored-by: tianqilin.99 <tianqilin.99@bytedance.com>
Co-authored-by: TianQiLin666666 <1834987979@qq.com>
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>

506c4928

17 Jun, 2025 2 commits
- fix: only enable flash_attn test on sm80 sm90 (#7289) · 0650e517
  Yineng Zhang authored Jun 17, 2025
  
  0650e517
- Fix AWQ Dequant and Weight Loading of deepseek v2 (#6842) · 3eb4a800
  AniZpZ authored Jun 18, 2025
  
  3eb4a800
16 Jun, 2025 1 commit
- Fix sampling for speculative decoding & simplify kernels (#7207) · cfceb83d
  Lianmin Zheng authored Jun 16, 2025
  
  cfceb83d
14 Jun, 2025 1 commit
- [fix] fix cutlass_mla_backend with cuda_graph and add sm_scale for sgl-kernel cutlass_mla (#7184) · ab1a4fa5
  JieXin Liang authored Jun 15, 2025
  
  ab1a4fa5
13 Jun, 2025 2 commits
- Support new DeepGEMM format in per token group quant (#7146) · 5c66c442
  fzyzcjy authored Jun 13, 2025
  
  5c66c442
- Remove 200us slow concat kernel (part 1: kernel) (#7145) · aa46ed34
  fzyzcjy authored Jun 13, 2025
  
  aa46ed34
12 Jun, 2025 1 commit
- [sgl-kernel] Add cuda kernel for moe_ep_silu_and_mul (#6919) · 84727a51
  Yuan Luo authored Jun 12, 2025
```
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
```
  84727a51
09 Jun, 2025 1 commit
- [perf][sgl-kernel] extend cutlass_mla_decode to support num_head < 128 (#6929) · 18efb5e8
  JieXin Liang authored Jun 09, 2025
  
  18efb5e8
05 Jun, 2025 2 commits
- [EP] Add cuda kernel for moe_ep_post_reorder (#6837) · 43baba64
  Yuan Luo authored Jun 05, 2025
```
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
```
  43baba64
- support 1 shot allreduce in 1-node and 2-node using mscclpp (#6277) · 8e3797be
  zyksir authored Jun 05, 2025
  
  8e3797be
04 Jun, 2025 3 commits
- Set `num_fused_shared_experts` as `num_shared_experts` when shared_experts... · 81964328
  Cheng Wan authored Jun 04, 2025
```
Set `num_fused_shared_experts` as `num_shared_experts` when shared_experts fusion is not disabled (#6736)
```
  81964328
- fix ep_moe_reorder kernel bugs (#6858) · bd75690f
  Xiaoyu Zhang authored Jun 04, 2025
```
Co-authored-by: JieXin Liang <Alcanderian@users.noreply.github.com>
```
  bd75690f
- [Refactor] Rename `n_share_experts_fusion` as `num_fused_shared_experts` (#6735) · 8a548052
  Cheng Wan authored Jun 03, 2025
  
  8a548052
26 May, 2025 1 commit
- [FA][Test] Fix Sparse FA test (#6306) · 006ead9d
  Brayden Zhong authored May 26, 2025
  
  006ead9d
22 May, 2025 2 commits
- Add main for merge state tests (#6492) · 121f92c5
  Yuan Luo authored May 22, 2025
```
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
```
  121f92c5
- [1/2] Support Qserve (#6457) · 4d643f6c
  HandH1998 authored May 22, 2025
```
Co-authored-by: yych0745 <1398089567@qq.com>
Co-authored-by: sleepcoo <sleepcoo@gmail.com>
```
  4d643f6c
16 May, 2025 1 commit
- [2/2] Add python wrapper for CUTLASS FP8 Blockscale MoE Kernel. (#5694) · 6fc93575
  Elfie Guo authored May 16, 2025
  
  6fc93575
12 May, 2025 2 commits
- Revert "fix some typos" (#6244) · e8e18dcd
  Lianmin Zheng authored May 12, 2025
  
  e8e18dcd
- fix some typos (#6209) · d738ab52
  applesaucethebun authored May 12, 2025
```
Co-authored-by: Brayden Zhong <b8zhong@uwaterloo.ca>
```
  d738ab52
11 May, 2025 1 commit
- Add typo checker in pre-commit (#6179) · 2ce87935
  applesaucethebun authored May 11, 2025
```
Co-authored-by: Brayden Zhong <b8zhong@uwaterloo.ca>
```
  2ce87935
09 May, 2025 1 commit
- Cutlass MLA: Disable split kv due to https://github.com/NVIDIA/cutlass/issues/2274 (#6101) · 0ab3f437
  Trevor Morris authored May 08, 2025
  
  0ab3f437
30 Apr, 2025 1 commit
- [Feat] Scale up fa3 kernel to sm8x arch (#5912) · 08acdb5c
  PGFLMG authored May 01, 2025
```
Co-authored-by: zhyncs <me@zhyncs.com>
```
  08acdb5c
28 Apr, 2025 1 commit
- [Feat] QWen-1M context support[1/2]: Update block sparse attention backend utils kernel (#5847) · ee71ed8a
  PGFLMG authored Apr 29, 2025
```
Co-authored-by: sighingnow <sighingnow@gmail.com>
```
  ee71ed8a
23 Apr, 2025 2 commits
- fix sgl-kernel unit tests (#5666) · 15fabcc0
  Yineng Zhang authored Apr 23, 2025
  
  15fabcc0
- [1/2] Add FP8 Blockscale MoE CUTLASS kernel for Blackwell (#5281) · e62c4955
  Elfie Guo authored Apr 22, 2025
  
  e62c4955
18 Apr, 2025 1 commit
- Sgl kernel fused_moe_gate support n_shared_experts (#5440) · 8e09b370
  Xiaoyu Zhang authored Apr 18, 2025
  
  8e09b370
17 Apr, 2025 2 commits
- [Feat] Update sgl-kernel flashinfer to latest main version (#5500) · c08a717c
  PGFLMG authored Apr 18, 2025
```
Co-authored-by: zhyncs <me@zhyncs.com>
```
  c08a717c
- Add test for flash_attn_varlen_func kernel (#5484) · 81c89111
  Baizhou Zhang authored Apr 17, 2025
  
  81c89111
15 Apr, 2025 2 commits
- BLackwell cutlass mla: Add check for bad page size/block num combinations (#5431) · e8f62b20
  Trevor Morris authored Apr 15, 2025
  
  e8f62b20
- kernel: support slightly faster merge_state_v2 cuda kernel (#5381) · 388e15c0
  DefTruth authored Apr 15, 2025
  
  388e15c0
13 Apr, 2025 1 commit
- feat: adapt merge_state (#5337) · b62e7e99
  Yineng Zhang authored Apr 12, 2025
  
  b62e7e99
12 Apr, 2025 3 commits
- [Feat] Add sparse attn to sgl-kernel (#5327) · 4879e50c
  PGFLMG authored Apr 13, 2025
  
  4879e50c
- update variable naming and comments for rocm (#5299) · 3c9740d2
  Zhaoyi Li authored Apr 12, 2025
  
  3c9740d2
- Blackwell Cutlass MLA kernel (#5142) · f65b8d5c
  Trevor Morris authored Apr 11, 2025
  
  f65b8d5c
11 Apr, 2025 2 commits
- fix: remove cublas_grouped_gemm (#5307) · 136b8e6a
  Yineng Zhang authored Apr 11, 2025
  
  136b8e6a
- fix: use fa3 unit test on hopper only (#5304) · c1dd773c
  Yineng Zhang authored Apr 11, 2025
  
  c1dd773c
10 Apr, 2025 2 commits
- [Misc] Clean sgl-kernel test (#5216) · ed01b451
  PGFLMG authored Apr 11, 2025
  
  ed01b451
- reduce moe_align_block_size_kernel small batch mode overhead (#5086) · f730362e
  Xiaoyu Zhang authored Apr 10, 2025
  
  f730362e