Commits · e1eefa4c40fc5b28bd7e83b6596bb5d2f420fd92 · OpenDAS / vllm_cscc

21 Nov, 2025 2 commits

[Bug] Fix torch warning of tf32 usage (#29112) · e1eefa4c
Wentao Ye authored Nov 20, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
e1eefa4c

[Feature] Shared Experts Overlap with FI deepgemm swap kernel, 2.2% throughput... · df44df01

Wentao Ye authored Nov 20, 2025


[Feature] Shared Experts Overlap with FI deepgemm swap kernel, 2.2% throughput improvement and 3.6% TTFT improvement (#28879)
Signed-off-by: yewentao256 <zhyanwentao@126.com>

df44df01

20 Nov, 2025 3 commits
- Add truncate arg to yarn to match openai implementation of gpt-oss (#28244) · 6eb745d9
  Anna Shors authored Nov 20, 2025
```
Signed-off-by: ashors1 <ashors@nvidia.com>
Co-authored-by: Chen Zhang <zhangch99@outlook.com>
```
  6eb745d9
- [Bug] Fix torch dynamo warning Dynamo detected a call to a `functools.lru_cache` (#29038) · 2c52c7fd
  Wentao Ye authored Nov 20, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  2c52c7fd
- Consolidate Nvidia ModelOpt quant config handling for all quantization methods (#28076) · a8c53682
  Shengliang Xu authored Nov 19, 2025
```
Signed-off-by: Shengliang Xu <shengliangx@nvidia.com>
```
  a8c53682
19 Nov, 2025 16 commits

[Refactor] Optimize `select_experts` (#28069) · 5031cd5d
Wentao Ye authored Nov 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
5031cd5d
[Feature] EPLB on Qwen3VLMoe and CompressedTensorsWNA16MoEMethod (#28849) · 8e38e998
JartX authored Nov 20, 2025

8e38e998
[Bugfix] Move flashinfer kernel check into ```__init__``` function of ```FusedMoE``` (#29018) · cb0a7b4b
Max Hu authored Nov 19, 2025
```
Signed-off-by: Max Hu <hyoung2991@gmail.com>
```
cb0a7b4b
[DeepSeek] Fix DeepSeek V3.2 Rope Embedding (#28968) · 88f5b19f
Yongye Zhu authored Nov 19, 2025
```
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
```
88f5b19f

[MoE] Nvfp4 Masked Gemm: Add flashinfer grouped_gemm_nt_masked (#25990) · 613abb50

Shu Wang authored Nov 19, 2025


Signed-off-by: Shu Wang. <shuw@nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

613abb50

[Bug] Fix Batch Invariant MLA test (#28967) · 1607e664
Wentao Ye authored Nov 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
1607e664

[Feature] Prefill Context Parallel (PCP) basic support (#28718) · 2fd893b4

Qiu authored Nov 20, 2025


Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
Signed-off-by: FENP <yuanyongjie.yyj@antgroup.com>
Signed-off-by: LookAround <lixushi@huawei.com>
Signed-off-by: Jingchun Gao <gaojingchun1@huawei.com>
Signed-off-by: zhenwenqi2024 <zhenwenqi_2022@qq.com>
Co-authored-by: FENP <yuanyongjie.yyj@antgroup.com>
Co-authored-by: LookAround <lixushi@huawei.com>
Co-authored-by: Jingchun Gao <gaojingchun1@huawei.com>
Co-authored-by: zhenwenqi2024 <zhenwenqi_2022@qq.com>
Co-authored-by: Jingchun Gao <63247409+gjc0824@users.noreply.github.com>

2fd893b4

[Bugfix] Fix precision corruption when shared_experts_stream=None (#28942) · 9d2d5612
杰兮 authored Nov 20, 2025
```
Signed-off-by: zhyajie <yajizhan@amd.com>
Co-authored-by: zhyajie <yajizhan@amd.com>
```
9d2d5612
[Kernels] Improve H200 Fused MoE Config (#28992) · fe69f331
Robert Shaw authored Nov 19, 2025
```
Signed-off-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
```
fe69f331
Update `rope_scaling` to `rope_parameters` in preparation for Transformers v5 (#28542) · a8b70304
Harry Mellor authored Nov 19, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
a8b70304

[Model][Mamba] Add selector for mamba attention backend and make it pluggable... · d44e9df7

Shanshan Shen authored Nov 20, 2025


[Model][Mamba] Add selector for mamba attention backend and make it pluggable for other device (#26487)
Signed-off-by: shen-shanshan <467638484@qq.com>

d44e9df7

[Feat][Perf] Enable deepep-low-latency with round-robin expert placement. (#28449) · da2f6800

Chen Bruce authored Nov 19, 2025


Signed-off-by: bruceszchen <bruceszchen@tencent.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

da2f6800

[Bugfix] Fix FusedMoEModularKernel for triton backend (#28913) · 468a8d72
Xin Yang authored Nov 18, 2025
```
Signed-off-by: Xin Yang <xyangx@amazon.com>
```
468a8d72
[CPU] Refactor CPU WNA16 (#28826) · 20852c8f
Li, Jiang authored Nov 19, 2025
```
Signed-off-by: jiang1.li <jiang1.li@intel.com>
```
20852c8f

[Hybrid][torch.compile] Refactor mamba2 forward to avoid obscuring linear... · 1395461f

tomeras91 authored Nov 19, 2025


[Hybrid][torch.compile] Refactor mamba2 forward to avoid obscuring linear projections under custom op (#28587)
Signed-off-by: Tomer Asida <57313761+tomeras91@users.noreply.github.com>

1395461f

[Build] Add OpenAI triton_kernels (#28788) · 9912b8cc

Varun Sundar Rabindranath authored Nov 18, 2025


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

9912b8cc

18 Nov, 2025 5 commits

[Models] Replace all `nn.Conv2d` with vLLM's Conv2dLayer (#28842) · e4bb2684
Isotr0py authored Nov 19, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
e4bb2684

[Model] Add Gemma3 GGUF multimodal support (#27772) · c2612371

Luciano Martins authored Nov 18, 2025

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

c2612371

[Model][Perf] Use cos and sin cache in QwenVL (#28798) · b9489f51
Canlin Guo authored Nov 18, 2025
```
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
```
b9489f51
[Refactor] Remove Unused Func in Batch Invariant (#28881) · 3ddcf460
Wentao Ye authored Nov 17, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
3ddcf460

[ROCm][Quantization] add apply_vllm_mapper in quark config for models like gpt-oss (#28638) · d0a73620

xuebwang-amd authored Nov 18, 2025


Signed-off-by: xuebwang-amd <xuebwang@amd.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

d0a73620

17 Nov, 2025 2 commits
- [Bugfix] Fix GPT-OSS on AMD after #28603 (#28816) · f8b19c0f
  Zhewen Li authored Nov 17, 2025
```
Signed-off-by: zhewenli <zhewenli@meta.com>
```
  f8b19c0f
- [Performance][Fix] update nvfp4 code to support renorm routing (#28569) · 561253b3
  jiahanc authored Nov 16, 2025
```
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
```
  561253b3
16 Nov, 2025 1 commit
- Feature: Support Relu2 in FusedMoE fp8 cutlass path (#27261) · 03ee4811
  amirkl94 authored Nov 16, 2025
  
  03ee4811
15 Nov, 2025 2 commits
- [Kernel][Moe Configs] llama4 maverick fp8 moe config tp8 on mi325 (#28709) · 1ec978c2
  Zhewen Li authored Nov 15, 2025
```
Signed-off-by: Zhewen Li <zhewenli@meta.com>
```
  1ec978c2
- [Performance][DeepGEMM] Estimate expected_m (#28694) · 6965ef43
  Varun Sundar Rabindranath authored Nov 15, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  6965ef43
14 Nov, 2025 8 commits
- [Hybrid] [Kernel] Fix chunk scan kernel when BLOCK_SIZE_DSTATE > 128 (#28295) · e0c910bb
  Thomas Parnell authored Nov 14, 2025
```
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
```
  e0c910bb
- [Bugfix] Fix incorrect use of hidden_states for shared_experts due to... · e5c78956
  Alexander Matveev authored Nov 14, 2025
```
[Bugfix] Fix incorrect use of hidden_states for shared_experts due to do_naive_dispatch_combine (#28740)
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
```
  e5c78956
- [BugFix] Fix misprint introduced by modular_kernel refactoring. (#28728) · fd455508
  Andrey Khalyavin authored Nov 14, 2025
```
Signed-off-by: Andrey Khalyavin <halyavin@yandex-team.ru>
```
  fd455508
- [Bugfix] [ROCm] [AITER]: Fix aiter block quant not compatible with torch compile dynamo (#28716) · a425dc25
  TJian authored Nov 14, 2025
```
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
```
  a425dc25
- [Kernels] Enable FlashInfer FP8 Blockscale on SM90 (for TEP DSR1) (#27134) · 3f8a8740
  Duncan Moss authored Nov 14, 2025
```
Signed-off-by: Duncan Moss <djm.moss@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>
```
  3f8a8740
- [Model][MM] Extract conv layer as CustomOp (#28455) · 41b92f7d
  Shanshan Shen authored Nov 14, 2025
```
Signed-off-by: shen-shanshan <467638484@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  41b92f7d
- [Misc] add ignore mapper for quark quantization (#28275) · 0b254989
  haoyangli-amd authored Nov 14, 2025
```
Signed-off-by: Haoyang Li <lihaoyang0109@gmail.com>
```
  0b254989
- [quantization][config] enable override existing quant_config (#28510) · 4d5943bd
  Hank_ authored Nov 14, 2025
```
Signed-off-by: Hank <hcc.mayday@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
```
  4d5943bd
13 Nov, 2025 1 commit

[Performance][B200] silu_mul_quant: pack scales in int32 (#28358) · fe1cd770

Varun Sundar Rabindranath authored Nov 13, 2025


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

fe1cd770