Commits · de2b78305f4e8fa07e0b8b2a4fb6ac03ab0dde20 · OpenDAS / vllm_cscc

09 Nov, 2025 2 commits
- [ROCm] Add env to enable/disable aiter triton gemm (#28321) · de2b7830
  Yong Hoon Shin authored Nov 08, 2025
```
Signed-off-by: Yong Hoon Shin <yhshin@meta.com>
```
  de2b7830
- [Misc] fix typo and add detailed log (#28178) · e5e9067e
  Ning Xie authored Nov 09, 2025
```
Signed-off-by: Andy Xie <andy.xning@gmail.com>
```
  e5e9067e
08 Nov, 2025 15 commits

[Performance][gpt-oss] Revert gpt-oss max cudagraph size to 1024 (#28345) · 404d7a9d
Mohammad Miadh Angkad authored Nov 09, 2025
```
Signed-off-by: Mohammad Miadh Angkad <MAngkad.BSDSBA2027@aim.edu>
```
404d7a9d
[Feat] Drop-in Torch CUDA Profiler (#27841) · 975676d1
Benjamin Chislett authored Nov 08, 2025
```
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
```
975676d1

[DCP] Support dcp kv_cache interleave size > 1 (#26696) · 2108a571

zhangsicheng5 authored Nov 09, 2025


Signed-off-by: zhangsicheng5 <zhangsicheng5@huawei.com>
Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
Signed-off-by: Qiu <qiuchunshuo@huawei.com>
Co-authored-by: QiuChunshuo <qiuchunshuo@huawei.com>

2108a571

[Bugfix] Spec decode + structured output + spec model max len edge case (#28298) · 47604137
Andy Lo authored Nov 08, 2025
```
Signed-off-by: Andy Lo <andy@mistral.ai>
```
47604137
[Bugfix] Update device name for H200 detection (#28349) · 26990d25
Robert Shaw authored Nov 08, 2025
```
Signed-off-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
```
26990d25
`reasoning_content` -> `reasoning` (#27752) · d9ab1ad9
Harry Mellor authored Nov 08, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
d9ab1ad9
[Attention] Remove max cudagraph size limit of 992 (#27840) · 608bb144
22quinn authored Nov 07, 2025
```
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
```
608bb144
[flashinfer][fix] do not check nvcc availability when using pre-downloaded cubins (#27990) · 4a36681f
Xiaozhu Meng authored Nov 07, 2025
```
Signed-off-by: Xiaozhu <mxz297@gmail.com>
Co-authored-by: Lu Fang <30275821+houseroad@users.noreply.github.com>
```
4a36681f

[Model] Consolidate Deepseek-MoE implementation with DeepSeek-v2 (#28101) · 934a9c3b

Isotr0py authored Nov 08, 2025


Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>

934a9c3b

[bugfix] support eagle with lora cudagraph specialization (#28318) · 70af44fd
gnovack authored Nov 07, 2025
```
Signed-off-by: gnovack <gnovack@amazon.com>
```
70af44fd

Bump arctic-inference requirement (#28174) · 781f5ebf

Aurick Qiao authored Nov 07, 2025


Co-authored-by: Aurick Qiao <aurick.qiao@snowflake.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>

781f5ebf

[Perf][DeepSeek] Add sigmoid+bias fusion to fused_grouped_topk from TRTLLM (#28124) · 08525276

Michael Goin authored Nov 08, 2025

Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>

08525276

[Bugfix][LoRA][Spec Decode] Support LoRA with speculative decoding (#21068) · d0c77920

Xiaohong (Sean) Chen authored Nov 07, 2025

Signed-off-by: Sean Chen <xiaohong_chen1991@hotmail.com>
Signed-off-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Danielle Robinson <dcmaddix@gmail.com>
Co-authored-by: Haipeng Li <li2haipeng@gmail.com>
Co-authored-by: li2haipeng <44383182+li2haipeng@users.noreply.github.com>

d0c77920

remove resolve_op_overloads and use splitting_ops directly (#28081) · b158df28
Boyuan Feng authored Nov 07, 2025
```
Signed-off-by: Boyuan Feng <boyuan@meta.com>
```
b158df28
[XPU] Enable Expert parallel for MoE models (#28263) · 1aaecda0
Kunshang Ji authored Nov 08, 2025
```
Signed-off-by: Yan Ma <yan.ma@intel.com>
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
```
1aaecda0

07 Nov, 2025 19 commits
- [PerfFix] Avoid separate thread for MP executor shm spin (take 2) (#28319) · 67a2da89
  Nick Hill authored Nov 07, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  67a2da89
- [Core] Rework handling of async scheduling config (#28250) · da786e33
  Nick Hill authored Nov 07, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  da786e33
- [Bugfix] Fix and add tests for GptOss reasoning parser (#28000) · 18903216
  Benjamin Chislett authored Nov 07, 2025
```
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
```
  18903216
- Revert "[PerfFix] Avoid separate thread for MP executor shm spin (#28012)" (#28289) · 68a72a5c
  Nicolò Lucchesi authored Nov 07, 2025
```
Signed-off-by: NickLucche <nlucches@redhat.com>
```
  68a72a5c
- [Log] update shm wait time msg (#28255) · 0f872b79
  Boyuan Feng authored Nov 07, 2025
  
  0f872b79
- [Feature] Default `ignore_eos` True for `random` dataset (#28227) · 4b1ff132
  Wentao Ye authored Nov 07, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  4b1ff132
- [CLI] add --max-tokens to `vllm complete` (#28109) · e0d6b4a8
  Iceber Gu authored Nov 07, 2025
```
Signed-off-by: Iceber Gu <caiwei95@hotmail.com>
```
  e0d6b4a8
- [Bugfix] Use latency MOE backend as default for Flashinfer and other misc fixes (#27439) · 72b1c2ae
  Pavani Majety authored Nov 07, 2025
```
Signed-off-by: Pavani Majety <pmajety@nvidia.com>
```
  72b1c2ae
- [Core][MM] Add mechanism to configure multimodal fields which should stay on CPU (#28168) · e0919f33
  Lukas Geiger authored Nov 07, 2025
```
Signed-off-by: Lukas Geiger <lukas.geiger94@gmail.com>
```
  e0919f33
- [fix] Revert "fixing mm placeholder replacement issue with gemma3" (#28285) · 8e19d470
  Kevin H. Luu authored Nov 07, 2025
```
Signed-off-by: Kevin H. Luu <khluu000@gmail.com>
```
  8e19d470
- [Misc][Model][Refactor] Pass the prefix into Linear layers (#28259) · 1958bda9
  Mengqing Cao authored Nov 07, 2025
```
Signed-off-by: MengqingCao <cmq0113@163.com>
```
  1958bda9
- [CPU]Avoid repeated random sample compile (#28260) · 7bdb42b2
  Zhang Xiangze authored Nov 07, 2025
```
Signed-off-by: Zhang Xiangze <Xiangze.Zhang@arm.com>
```
  7bdb42b2
- [FixBug]Aeala/ShareGPT_Vicuna_unfiltered marked as multimodal benchmark (#28265) · 315068eb
  汪志鹏 authored Nov 07, 2025
```
Signed-off-by: princepride <wangzhipeng628@gmail.com>
```
  315068eb
- [Perf] Introduce FlattenLogprobs to store logprobs results to reduce GC overhead (#28171) · ccd98b59
  Jialin Ouyang authored Nov 07, 2025
```
Signed-off-by: Jialin Ouyang <Jialin.Ouyang@gmail.com>
```
  ccd98b59
- [Kernel] LoRA triton kernels support PDL (#27402) · 21b82f4e
  Jee Jee Li authored Nov 07, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  21b82f4e
- [Bug] Fix missing token_ids for reasoning parser models in chat completions #28246 (#28256) · 9da9208b
  baonudesifeizhai authored Nov 07, 2025
  
  9da9208b
- [amd][gptoss] Perf gain because of block alignment (#28024) · 11fd69dd
  smit kadvani authored Nov 06, 2025
```
Signed-off-by: Smit Kadvani <smit.kadvani@gmail.com>
Co-authored-by: Smit Shaileshbhai Kadvani <kadvani@meta.com>
```
  11fd69dd
- Fix issues from #28242 (#28257) · c0a4b95d
  Harry Mellor authored Nov 06, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  c0a4b95d
- [Multimodal][torch.compile] Add compilation config field for turning off ViT/MM compile (#28242) · 4bf56c79
  Lucas Kabela authored Nov 06, 2025
```
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
```
  4bf56c79
06 Nov, 2025 4 commits

Speed up mm processor kwargs per request by spliting dynamic and static kwargs (#26483) · 59b453ea

Junhong Liu authored Nov 07, 2025


Signed-off-by: Junhong <liujunhong11@huawei.com>
Signed-off-by: Junhong Liu <98734602+LJH-LBJ@users.noreply.github.com>
Co-authored-by: Junhong <liujunhong11@huawei.com>

59b453ea

[BugFix] Fix FusedMoELoRA + ModularKernel Integration (#28237) · ca6f755d

Varun Sundar Rabindranath authored Nov 06, 2025


Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>

ca6f755d

[ROCm] triton fp8 kernel (#27058) · 449de900

Aleksandr Malyshev authored Nov 06, 2025

Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Gregory Shtrasberg <156009573+gshtras@users.noreply.github.com>

449de900

[Chore] eliminate duplicated and unconditional object serialization in... · d4aa65c9

Vico Chu authored Nov 07, 2025


[Chore] eliminate duplicated and unconditional object serialization in anthropic messages api (#27792)
Signed-off-by: Vico Chu <vico24826@gmail.com>

d4aa65c9