Commits · 0b8c5721f1ed373d75c4bcc8cbdfad0de8c1770a · change / sglang

"vscode:/vscode.git/clone" did not exist on "1e2cf2b541c7b22dbf5f0d5bf6da78d93eaa3648"

06 Sep, 2025 6 commits
- [HiStorage] Remove delete and clear as necessary methods (#10039) · 0b8c5721
  Zhiqiang Xie authored Sep 05, 2025
  
  0b8c5721
- Add lora_path argument to bench_multiturn.py (#10092) · beac202b
  Baizhou Zhang authored Sep 05, 2025
  
  beac202b
- [router] Introduce router integration tests (#10086) · 21b9a4b4
  Keyang Ru authored Sep 05, 2025
  
  21b9a4b4
- [router] move to mcp sdk instead (#10057) · db37422c
  Simo Lin authored Sep 05, 2025
  
  db37422c
- support Llama4 with non uniformed intermediate size across layers for… (#10047) · ab62b135
  gongwei-130 authored Sep 05, 2025
  
  ab62b135
- [Minor] Refactors KV memory pool (#9842) · 273b2834
  Xinyuan Tong authored Sep 06, 2025
  
  273b2834
05 Sep, 2025 25 commits
- Add storage read/write bandwidth logs to monitor kvcache performance (#9965) · f84db115
  pansicheng authored Sep 06, 2025
```
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  f84db115
- Update wave-lang to 3.7.0 and unify Wave kernel buffer options (#10069) · efb0de2c
  jacky.cheng authored Sep 06, 2025
  
  efb0de2c
- [Bug Fix] Fix Glm4vVisionBlock norm (#9884) · 0f6ac5e2
  Adam Yanxiao Zhao authored Sep 06, 2025
  
  0f6ac5e2
- Update flashinfer to 0.3.1 for B300 support (#10087) · 29850900
  hlu1 authored Sep 05, 2025
```
Signed-off-by: Hao Lu <14827759+hlu1@users.noreply.github.com>
```
  29850900
- [bugfix]: use correct cache location for cross attention in torch native backend (#8622) · e678cc71
  Mahmoud Ashraf authored Sep 05, 2025
  
  e678cc71
- enable aiter gemm_a8w8_bpreshuffle for ptpc gemm (#8555) · 4efe844a
  Morpheus Guo authored Sep 06, 2025
  
  4efe844a
- [router] add rust cache in benchmark ci (#10080) · bde73ee4
  Simo Lin authored Sep 05, 2025
  
  bde73ee4
- [router] add rust cache for rust unit test (#10079) · 4f0e28d7
  Keyang Ru authored Sep 05, 2025
  
  4f0e28d7
- [router] add py binding unit tests to coverage 80% (#10043) · 045ab92d
  Keyang Ru authored Sep 05, 2025
  
  045ab92d
- Support copying tensor from cpu to gpu without using copy engines (#10007) · bd7f8821
  fzyzcjy authored Sep 05, 2025
  
  bd7f8821
- Tiny let DeepGEMM scale checks cover more cases (#7182) · 5e5c30d9
  fzyzcjy authored Sep 05, 2025
```
Co-authored-by: Yineng Zhang <me@zhyncs.com>
```
  5e5c30d9
- Fix and enhance dumper (#8725) · 9f00ec44
  fzyzcjy authored Sep 05, 2025
  
  9f00ec44
- Support simple evals in text comparator (#8867) · 8e85ee88
  fzyzcjy authored Sep 05, 2025
  
  8e85ee88
- Forbid DeepEP racing condition when too many tokens (#9567) · adf73175
  fzyzcjy authored Sep 05, 2025
  
  adf73175
- [Fix] Add speculative_draft_model_revision to server_args (#5255) · 13705dae
  DevashishLal-CB authored Sep 05, 2025
```
Signed-off-by: Devashish Lal <devashish@rivosinc.com>
```
  13705dae
- Tiny support setting numa nodes for different ranks (#10006) · df97b31f
  fzyzcjy authored Sep 05, 2025
  
  df97b31f
- [1/2] Optimizations and refactors about quant kernel (#9534) · 339f8eef
  fzyzcjy authored Sep 05, 2025
  
  339f8eef
- Fix typo in scheduler (#9934) · afd9f2f5
  limingshu authored Sep 05, 2025
  
  afd9f2f5
- [Vulnerability]feat(conn): set bootstrap server host (#9931) · f40038fb
  Jimmy authored Sep 05, 2025
  
  f40038fb
- Integrate trtllm ragged attention for prefill self-attention (#9801) · bebd0576
  Elfie Guo authored Sep 05, 2025
  
  bebd0576
- fix MultiTokenizerWrapper name (#10049) · f9836660
  Huang Long authored Sep 05, 2025
```
Signed-off-by: huanglong <huanglong@linux.alibaba.com>
```
  f9836660
- [router] fix release workflow to include protobuf (#10055) · 8b3b995a
  Chang Su authored Sep 04, 2025
  
  8b3b995a
- Simplify `Router` arguments passing and build it in docker image (#9964) · 6e95f5e5
  Liangsheng Yin authored Sep 05, 2025
  
  6e95f5e5
- fix: update gb200 dep (#10052) · 0e9387a9
  Yineng Zhang authored Sep 04, 2025
  
  0e9387a9
- chore: bump v0.5.2rc2 (#10050) · fa9c82d3
  Yineng Zhang authored Sep 04, 2025
  
  fa9c82d3
04 Sep, 2025 9 commits
- Fix accuracy drop of dsv3 run in dp enablement (#8677) · 918e3d4c
  kk authored Sep 05, 2025
```
Co-authored-by: wunhuang <wunhuang@amd.com>
```
  918e3d4c
- Optimized deepseek-v3/r1 model performance on mxfp4 run (#10008) · e9697374
  kk authored Sep 05, 2025
```
Co-authored-by: wunhuang <wunhuang@amd.com>
Co-authored-by: HAI <hixiao@gmail.com>
Co-authored-by: Hubert Lu <55214931+hubertlu-tw@users.noreply.github.com>
```
  e9697374
- [Hicache] Mooncake API Fix & Test, and Improved Readme (#9951) · 93088b69
  ykwd authored Sep 05, 2025
```
Co-authored-by: Teng Ma <sima.mt@alibaba-inc.com>
```
  93088b69
- Save memory for expert model parallel (#9957) · 453511ac
  Cheng Wan authored Sep 04, 2025
  
  453511ac
- fix 3fs zerocopy (#9938) · d0730487
  pansicheng authored Sep 05, 2025
```
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  d0730487
- metrics: support customer buckets for prompt/generation_tokens_histogram (#9634) · b32ab070
  Yingchun Lai authored Sep 04, 2025
  
  b32ab070
- [Doc] Fix SGLang tool parser doc (#9886) · 75ee0011
  Huapeng Zhou authored Sep 04, 2025
  
  75ee0011
- Optimize Qwen3-moe model by using flashinfer fused allreduce (#9973) · ec15c836
  Yuan Luo authored Sep 04, 2025
```
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
```
  ec15c836
- feat(hicache): Add generic hicache ci e2e test and benchmark test (#9846) · 106c2b31
  hzh0425 authored Sep 04, 2025
```
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  106c2b31