Commits · 84c276d7ea00c5bcb6af21b8035d57735479e0ab · OpenDAS / vllm_cscc

27 Apr, 2026 2 commits

[Bugfix] Cap SWA/chunked-local runtime admission to startup pool-sizing bound (#40946) · 84c276d7

Dao007forever authored Apr 26, 2026


Signed-off-by: Dao Le <Dao007forever@gmail.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
(cherry picked from commit 7b1bc0a3eb01a6bc2650eda9970049f7825240d7)

84c276d7

[Feat] DeepSeek V4 Rebased (#40860) · 4d51588e

Yifan Qiao authored Apr 26, 2026


Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Signed-off-by: qizixi <zixi@inferact.ai>
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Yongye Zhu <yongye@inferact.ai>
Co-authored-by: Simon Mo <simon@inferact.ai>
Co-authored-by: Bugen Zhao <i@bugenzhao.com>
Co-authored-by: Giancarlo Delfin <gdelfin@inferact.ai>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roy Wang <yasong.wang@inferact.ai>
Co-authored-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Zhewen Li <jerven.vllm@gmail.com>
Co-authored-by: Zijing Liu <liuzijing2014@gmail.com>
Co-authored-by: khluu <khluu000@gmail.com>
Co-authored-by: qizixi <zixi@inferact.ai>
Co-authored-by: Zhewen Li <zhewenli@inferact.ai>

4d51588e

26 Apr, 2026 1 commit

[torch.compile]: Disable Sequence Parallelism (SP) for piecewise compilation (#38373) · 32e45636

Xinan Miao authored Apr 27, 2026

Signed-off-by: SouthWest7 <am1ao@qq.com>
Signed-off-by: Xinan Miao <1403572259@qq.com>
Co-authored-by: SouthWest7 <am1ao@qq.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: Wang Xingran <72983099+wangxingran222@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

32e45636

25 Apr, 2026 1 commit

[ROCm][Engine] Fix GPU memory leaks in engine shutdown and test workaround for... · 95995bbe

Andreas Karatzas authored Apr 25, 2026


[ROCm][Engine] Fix GPU memory leaks in engine shutdown and test workaround for async KV prefix cache reset (#38503)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>

95995bbe

24 Apr, 2026 8 commits

[Refactor] Remove unused dead code (#40640) · a474da28
Wentao Ye authored Apr 24, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
a474da28
[Bug] Fix GLM-5.1 running error on ROCm platform (#40763) · 095d2f87
qli88 authored Apr 24, 2026
```
Signed-off-by: Qiang Li <qiang.li2@amd.com>
```
095d2f87
[Refactor] Unify 2D/3D kernels in triton_unified_attention (#40631) · 914d0464
JartX authored Apr 24, 2026
```
Signed-off-by: JartX <sagformas@epdcenter.es>
```
914d0464
[Attention] use diff kv backend for mimo v2 flash (#40045) · e8ee2a78
Jiangyun Zhu authored Apr 24, 2026
```
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
```
e8ee2a78

[Model] Gemma4: add bidirectional vision attention for sliding layers with window guard (#40534) · 512f5221

Luciano Martins authored Apr 24, 2026


Signed-off-by: Luciano Martins <lucianommartins@users.noreply.github.com>
Signed-off-by: Luciano Martins <lucianomartins@google.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Luciano Martins <lucianommartins@users.noreply.github.com>
Co-authored-by: Isotr0py <2037008807@qq.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>

512f5221

[MRV2] Ensure warmup covers prefill path (#40746) · e9f331d7
Nick Hill authored Apr 23, 2026
```
Signed-off-by: Nick Hill <nickhill123@gmail.com>
```
e9f331d7

[Feat] Unified Synthetic Acceptance Rate for V1 and V2 (#40662) · 626daa20

Benjamin Chislett authored Apr 23, 2026

Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
Signed-off-by: Benjamin Chislett <chislett.ben@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

626daa20

[Core] Avoid seq_lens_cpu GPU->CPU sync (#40654) · fe85a92e
Nick Hill authored Apr 23, 2026
```
Signed-off-by: Nick Hill <nickhill123@gmail.com>
```
fe85a92e

23 Apr, 2026 5 commits

[Spec Decode] Move `SpecDecodeBaseProposer` out of `eagle.py` (#40732) · cde8d247
Matthew Bonanni authored Apr 23, 2026
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
cde8d247

[Bugfix] Fix workspace resize leaking reserved GPU memory (#39226) · 7ff65b19

czhu-cohere authored Apr 23, 2026


Signed-off-by: root <conway.zhu@cohere.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

7ff65b19

[DP][Ray] Pin DP control bundle to same node as first GPU bundle (#39167) · e9ba519f

shaharmor98 authored Apr 23, 2026

Signed-off-by: Shahar Mor <smor@nvidia.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

e9ba519f

[Bugfix] Make Attention Backend Auto-Selection Batch-Invariance-Aware (#40193) · b7a26050

Srreyansh Sethi authored Apr 23, 2026


Signed-off-by: Srreyansh Sethi <srreyansh.sethi@gmail.com>
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

b7a26050

[TurboQuant] enable FA3/FA4 for prefill paths (#40092) · fe9c3d6c

Julian Huang authored Apr 23, 2026


Signed-off-by: 墨楼 <huangzhilin.hzl@antgroup.com>
Co-authored-by: 墨楼 <huangzhilin.hzl@antgroup.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Codex <codex@openai.com>

fe9c3d6c

22 Apr, 2026 7 commits

[Refactor] Clean up log once `scope="local"` (#40540) · 8f87eb46

Wentao Ye authored Apr 22, 2026

Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

8f87eb46

FlexAttention non-causal support (#40394) · 29f64c5f

Fynn Schmitt-Ulms authored Apr 22, 2026


Signed-off-by: Fynn Schmitt-Ulms <fschmitt@redhat.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

29f64c5f

[Bugfix] LoRA for DeepSeek V3.2 (#35077) · a250f1bd

ℍ𝕠𝕝𝕝𝕠𝕨 𝕄𝕒𝕟 authored Apr 22, 2026


Signed-off-by: Hollow Man <hollowman@opensuse.org>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>

a250f1bd

[Model] Add block-local attention and YaRN for local layers to Gemma3 (#39823) · 12367487
philip-essential authored Apr 21, 2026
```
Signed-off-by: Philip Monk <169196560+philip-essential@users.noreply.github.com>
```
12367487

[kv_offload+HMA][8/N]: Support multi-group worker transfer (#38453) · aad88f84

Or Ozeri authored Apr 22, 2026


Signed-off-by: Or Ozeri <oro@il.ibm.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

aad88f84

[ROCm] Support non-causal attention in ROCM_ATTN (#40176) · 6d097697
Micah Williamson authored Apr 21, 2026
```
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
```
6d097697

[compile] mla + group fp8 fusion (#38877) · 4506319a

Carl Y authored Apr 21, 2026


Signed-off-by: Carl You <4531192+carlyou@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

4506319a

21 Apr, 2026 9 commits

[Startup][UX] Enable CUDAGraph memory profiling by default (#38284) · 96a85c57
Matthew Bonanni authored Apr 21, 2026
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
```
96a85c57
[Refactor] Remove unused param (#39750) · 1842447c
Wentao Ye authored Apr 21, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
1842447c

[Bugfix] Fix spec decode test failures on Blackwell (SM100+) (#39546) · 9f39b380

Rishi Puri authored Apr 21, 2026


Signed-off-by: Stefano Castagnetta <scastagnetta@nvidia.com>
Signed-off-by: Rishi Puri <puririshi98@berkeley.edu>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Stefano Castagnetta <scastagnetta@nvidia.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Benjamin Chislett <bchislett@nvidia.com>

9f39b380

[MRv2]fix: model accuracy regression caused by reusing the stale... · 9a6a66f3

Zijing Liu authored Apr 21, 2026


[MRv2]fix: model accuracy regression caused by reusing the stale last_sampled_tokens and draft_tokens (#39833)
Signed-off-by: Zijing Liu <liuzijing2014@gmail.com>

9a6a66f3

[Model Runner V2] Multiple prompt logprobs support (#39937) · 66cc3fa5

Wentao Ye authored Apr 21, 2026


Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>

66cc3fa5

[UX] Bump version in CG memory profiling log message (#40465) · ab5666eb
Matthew Bonanni authored Apr 21, 2026
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
ab5666eb

[MM][CG] Optimize default `max_frames_per_batch` auto-infer for ViT CUDA graph... · 936e0b79

Shanshan Shen authored Apr 21, 2026


[MM][CG] Optimize default `max_frames_per_batch` auto-infer for ViT CUDA graph video inference (#40445)
Signed-off-by: shen-shanshan <467638484@qq.com>

936e0b79

feat(multimodal): support externally processed mm_kwargs with cache injection (#39502) · 766cb65d

Kris Hung authored Apr 21, 2026


Signed-off-by: Krish Hung <krishung5@gmail.com>
Signed-off-by: krishung5 <krish@nvidia.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

766cb65d

[Feat] dflash support for ROCm (#39703) · f95c11a8
hangy-amd authored Apr 21, 2026
```
Signed-off-by: Hang Yang <hangy@amd.com>
```
f95c11a8

20 Apr, 2026 6 commits

[Misc][UX] Suppress confusing `num_gpu_blocks` log lines (#40402) · c075702e

Matthew Bonanni authored Apr 20, 2026


Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

c075702e

[Bug] Fix dcp error message (#40053) · b42e878e
Wentao Ye authored Apr 20, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
b42e878e

[ROCm][Feature] Enable AITER MLA attention backend to work with Eagle3... · 7243e02a

larryli2-amd authored Apr 20, 2026


[ROCm][Feature] Enable AITER MLA attention backend to work with Eagle3 speculative decoding on ROCm (#39616)
Signed-off-by: larryli2-amd <larryli2@amd.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>

7243e02a

[EPLB] Add nixl-based eplb communicator (#36276) · 50dd4cb4

Ilya Markov authored Apr 20, 2026


Signed-off-by: ilmarkov <markovilya197@gmail.com>
Signed-off-by: Markov Ilya <markovilya19@gmail.com>

50dd4cb4

[kv_offload+HMA][4/N]: Support sliding window lookup (#36645) · f774ba02
Or Ozeri authored Apr 20, 2026
```
Signed-off-by: Or Ozeri <oro@il.ibm.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
```
f774ba02
[XPU] [torch.compile] Skipping CUDA graph memory estimation to avoid startup errors. (#39977) · 4f4713f9
Chaojun Zhang authored Apr 20, 2026
```
Signed-off-by: chaojun-zhang <chaojun.zhang@intel.com>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
```
4f4713f9

19 Apr, 2026 1 commit
- [ROCm] Fix cu_seqlens_q off-by-one in AITER FA speculative decode path (#39120) · f150107e
  Andrew Barnes authored Apr 19, 2026
```
Signed-off-by: Bortlesboat <bortstheboat@gmail.com>
```
  f150107e