Commits · e4605d225e020154bc98efd15a05e11b83eaefb7 · OpenDAS / vllm_cscc

08 Dec, 2025 1 commit
- [Perf] Improve fp8 quant in mla; replace ReduceSum with ReduceScatterSum (#29795) · 1fb632fd
  Lain authored Dec 08, 2025
```
Signed-off-by: Siyuan Fu <siyuanf@nvidia.com>
```
  1fb632fd
07 Dec, 2025 1 commit
- [v1] Add PrefixLM support to FlexAttention backend (#27938) · b952f4d3
  Isotr0py authored Dec 07, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  b952f4d3
05 Dec, 2025 2 commits

[Attention][UX][1/N] Add AttentionConfig and change attention env vars to CLI arguments (#26315) · 66e674cd

Matthew Bonanni authored Dec 05, 2025

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Signed-off-by: Matthew Bonanni <mbonanni001@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>

66e674cd

[Bugfix] Correct num_q_heads on DCP for Flashinfer backends (#29487) · d698bb38

Jingchun Gao authored Dec 05, 2025


Signed-off-by: Jingchun Gao <gaojingchun1@huawei.com>
Signed-off-by: Jingchun Gao <63247409+gjc0824@users.noreply.github.com>
Co-authored-by: Jingchun Gao <gaojingchun1@huawei.com>

d698bb38

04 Dec, 2025 1 commit
- [ROCm][CI][Bugfix] Fixing the `Multi-Modal Models Test (Extended) 1` group (#30013) · e96a6a6d
  Andreas Karatzas authored Dec 04, 2025
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  e96a6a6d
02 Dec, 2025 1 commit
- [Attention][CUDAGraph] Remove CG padding from attention backends (#29352) · 1d93f116
  Matthew Bonanni authored Dec 02, 2025
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
  1d93f116
01 Dec, 2025 1 commit

[v1] Add real sliding window calculation to FlexAttention direct BlockMask building (#26015) · b95db244

Isotr0py authored Dec 01, 2025


Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: baonudesifeizhai <baonudesifeizhai@gmail.com>
Co-authored-by: baonudesifeizhai <baonudesifeizhai@gmail.com>

b95db244

30 Nov, 2025 2 commits
- [ROCm][Attention] Sliding window support for `AiterFlashAttentionBackend` (#29234) · 8c363ed6
  Pleaplusone authored Nov 30, 2025
```
Signed-off-by: ganyi <ygan@amd.com>
```
  8c363ed6
- Fix AttributeError about _use_fi_prefill (#29734) · 82c795d6
  Huamin Li authored Nov 29, 2025
```
Signed-off-by: Huamin Li <3ericli@gmail.com>
```
  82c795d6
29 Nov, 2025 1 commit
- [BugFix] Fix DBO failing with TypeError: 'NoneType' object is not iterable (#29698) · e23f665d
  Lucas Wilkinson authored Nov 28, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  e23f665d
28 Nov, 2025 1 commit
- bugfix: correct attn output with base 2 or e (#28840) · 9726e645
  Augusto Yao authored Nov 29, 2025
```
Signed-off-by: augusto.yjh <augusto.yjh@antgroup.com>
```
  9726e645
27 Nov, 2025 4 commits
- [BugFix] Fix new nightly failures (#29578) · be493e0b
  Lucas Wilkinson authored Nov 27, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  be493e0b
- [BugFix] Fix `plan` API Mismatch when using latest FlashInfer (#29426) · a5345bf4
  Andrii Skliar authored Nov 27, 2025
```
Signed-off-by: Andrii Skliar <askliar@askliar-mlt.client.nvidia.com>
Co-authored-by: Andrii Skliar <askliar@askliar-mlt.client.nvidia.com>
```
  a5345bf4
- [Attention] Update attention imports (#29540) · fc1d8be3
  Matthew Bonanni authored Nov 27, 2025
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
  fc1d8be3
- [Attention][Async] Eliminate `seq_lens_cpu` in FlashAttention metadata... · 77740191
  Matthew Bonanni authored Nov 26, 2025
```
[Attention][Async] Eliminate `seq_lens_cpu` in FlashAttention metadata building with DCP > 1 (#29449)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
  77740191
26 Nov, 2025 3 commits
- [Core] Refactor padding logic and pad for CUDA graphs before attention metadata building (#28579) · 56539cdd
  Lucas Wilkinson authored Nov 26, 2025
  
  56539cdd
- [Attention] Remove imports from `vllm/attention/__init__.py` (#29342) · 430dd4d9
  Matthew Bonanni authored Nov 26, 2025
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
  430dd4d9
- [Performance][MLA][ROCm] Remove redundant D2D copy in deepseek (#27457) · d9d342d2
  Pleaplusone authored Nov 26, 2025
```
Signed-off-by: ganyi <ygan@amd.com>
```
  d9d342d2
25 Nov, 2025 4 commits

[Core] Generalize Encoder-Decoder `seq_lens` computation to avoid Whisper... · 798e87db

Nicolò Lucchesi authored Nov 25, 2025


[Core] Generalize Encoder-Decoder `seq_lens` computation to avoid Whisper hardcoded logic   (#29268)
Signed-off-by: NickLucche <nlucches@redhat.com>
Co-authored-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>

798e87db

[Attention] add `_cudagraph_support` for linear attention (#28934) · 81db702e
Jiangyun Zhu authored Nov 25, 2025
```
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
```
81db702e

[ROCm][MLA] enable fp8 MLA decode on ROCm (#28032) · cb7214d8

gbyu-amd authored Nov 25, 2025


Signed-off-by: guanbao <gyu@amd.com>
Signed-off-by: Guanbao Yu <gyu@amd.com>
Signed-off-by: gbyu-amd <Guanbao.Yu@amd.com>
Co-authored-by: guanbao <gyu@amd.com>

cb7214d8

[Perf][Deepseek] optimize gather_and_maybe_dequant_cache kernel's perf for... · 77e10c9c

Pleaplusone authored Nov 25, 2025


[Perf][Deepseek] optimize gather_and_maybe_dequant_cache kernel's perf for extremely long sequence (#28029)
Signed-off-by: ganyi <ygan@amd.com>

77e10c9c

24 Nov, 2025 2 commits
- [Core] Deprecate `xformers` (#29262) · 0ff70821
  Roger Wang authored Nov 23, 2025
```
Signed-off-by: Roger Wang <hey@rogerw.io>
```
  0ff70821
- [ROCm] Support for Whisper v1 with Aiter Unified Attention and Aiter Flash Attention (#28376) · 5253f427
  tongqiu authored Nov 24, 2025
```
Signed-off-by: apinge <Tong.Qiu2@amd.com>
```
  5253f427
22 Nov, 2025 2 commits
- [perf][cpu] Accelerate paged attention GEMMs (QK, PV) on Arm CPUs with NEON (#29193) · 730bd353
  Fadi Arafeh authored Nov 22, 2025
```
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
```
  730bd353
- [Attention] Refactor FA `block_size` limitations to hybrid models only (#29084) · 066209a0
  Nicolò Lucchesi authored Nov 22, 2025
```
Signed-off-by: NickLucche <nlucches@redhat.com>
```
  066209a0
21 Nov, 2025 2 commits
- GPU Model Runner V2 (#25266) · 30b44a15
  Woosuk Kwon authored Nov 21, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  30b44a15
- fix cross attention (#28346) · fc9f821d
  who who who authored Nov 21, 2025
```
Signed-off-by: fsx950223 <fsx950223@outlook.com>
```
  fc9f821d
20 Nov, 2025 3 commits
- [KVConnector][Core] Support cross-layer KV blocks (#27743) · 64746471
  Or Ozeri authored Nov 20, 2025
```
Signed-off-by: Or Ozeri <oro@il.ibm.com>
```
  64746471
- [ROCm] Add AMD GPU support on Deepseek v3.2 and SparseMLA (#26670) · 06c20c99
  Pleaplusone authored Nov 20, 2025
```
Signed-off-by: ganyi <ygan@amd.com>
```
  06c20c99
- [AMD] Use Decoupled Kernel Block Size to Support AITER MLA block_size=1 (#27715) · 3fb0d909
  Qiang Zhang authored Nov 20, 2025
```
Signed-off-by: chiangzhang <chiangzhang@tencent.com>
```
  3fb0d909
19 Nov, 2025 4 commits

[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

Upstreaming aiter triton attention backend as a new backend (#28701) · ac10fd3c
Aleksandr Malyshev authored Nov 19, 2025
```
Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
```
ac10fd3c
[BugFix] Fix async-scheduling + FlashAttn MLA (#28990) · 48fc8b1e
Lucas Wilkinson authored Nov 19, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
48fc8b1e
[Attention] FlashAttention ViT support, make default backend (#28763) · 4c23690f
Matthew Bonanni authored Nov 18, 2025
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
4c23690f

17 Nov, 2025 1 commit
- [ROCm][Qwen3-32B] Fix AITER MHA accuracy issue cause by #25763 (#28670) · 60e089f0
  Xiake Sun authored Nov 17, 2025
```
Signed-off-by: Xiake Sun <xiake.sun@amd.com>
```
  60e089f0
14 Nov, 2025 3 commits

[BugFix] Fix FA3 IMA with FULL_AND_PIECEWISE and cascade attention (default) (#28702) · db56a599
Lucas Wilkinson authored Nov 14, 2025

db56a599
Fix KV sharing fast prefill with cudagraph enabled (#28537) · 9324e102
Yong Hoon Shin authored Nov 14, 2025
```
Signed-off-by: Yong Hoon Shin <yhshin@meta.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
9324e102

[DCP] Support Decode Context Parallel (DCP) for GQA with Flashinfer (#25438) · 4516d44b

Jingchun Gao authored Nov 14, 2025


Signed-off-by: gaojc <1055866782@qq.com>
Signed-off-by: Jingchun Gao <gaojingchun1@huawei.com>
Signed-off-by: Jingchun Gao <63247409+gjc0824@users.noreply.github.com>
Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
Co-authored-by: gaojingchun (A) <g00955623@china.huawei.com>
Co-authored-by: Jingchun Gao <gaojingchun1@huawei.com>
Co-authored-by: QiuChunshuo <qiuchunshuo@huawei.com>

4516d44b

13 Nov, 2025 1 commit

[bugfix] correct local_chunk_len for DCP in reorg_kvcache with long context (#28526) · 968060c1

Qiu authored Nov 14, 2025


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

968060c1