Commits · 69f30ae059e7623918637336457f1be42c8652c0 · OpenDAS / vllm_cscc · GitLab

01 Sep, 2025 1 commit
- add dca and sparse attention support on rocm · 12112a4d
  zhuwenwen authored Sep 01, 2025
  
  12112a4d
28 Aug, 2025 1 commit
- 优化 block_tables 的生成逻辑，增加对混合情况的检测，确保在存在空和非空块时正确计算最大块长度。 · 4932783d
  laibao authored Aug 28, 2025
  
  4932783d
15 Aug, 2025 1 commit
- 增加marlin对cache13的支持，以及新增flash mla的kvcache fp8的支持 · 8f25283a
  yangql authored Aug 15, 2025
  
  8f25283a
17 Jul, 2025 1 commit
- add VLLM_USE_PA_PRINT_PARAM to print fa-pa size · 1092a467
  zhuwenwen authored Jul 17, 2025
  
  1092a467
10 Jul, 2025 1 commit
- update bw vdim128 · d805c59c
  zhuwenwen authored Jul 10, 2025
  
  d805c59c
09 Jul, 2025 1 commit
- [fix]修复零消耗引入的cudagraph模式数据准备cpu耗时问题 · cd6239dc
  zhuwenwen authored Jul 09, 2025
  
  cd6239dc
06 Jul, 2025 2 commits
- Revert "[V0 deprecation] Remove V0 CPU/XPU/TPU backends (#20412)" · a5dd03c1
  simon-mo authored Jul 06, 2025
```
This reverts commit e202dd27.
```
  a5dd03c1
- [V0 deprecation] Remove V0 CPU/XPU/TPU backends (#20412) · e202dd27
  Woosuk Kwon authored Jul 06, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: jiang1.li <jiang1.li@intel.com>
Co-authored-by: Li, Jiang <jiang1.li@intel.com>
```
  e202dd27
03 Jul, 2025 1 commit
- update VLLM_USE_FLASH_ATTN_BACKEND to VLLM_USE_FLASH_ATTN_PA · e036115e
  zhuwenwen authored Jul 03, 2025
  
  e036115e
27 Jun, 2025 1 commit
- update cutlass fa and pa · 4ddb5447
  zhuwenwen authored Jun 27, 2025
  
  4ddb5447
23 Jun, 2025 1 commit
- update common.py · 2da6fd72
  zhuwenwen authored Jun 23, 2025
  
  2da6fd72
21 Jun, 2025 1 commit
- only K100_AI can use cutlass prefix cache · 14945681
  zhuwenwen authored Jun 21, 2025
  
  14945681
20 Jun, 2025 1 commit
- [Kernel] mark TorchSDPABackend swap_blocks NotImplementedError (#19749) · 71baf85a
  Ning Xie authored Jun 21, 2025
  
  71baf85a
17 Jun, 2025 1 commit
- [V1][Kernel] Flashinfer HND KV cache layout (#19280) · 4c8f64fa
  Nicolò Lucchesi authored Jun 17, 2025
```
Signed-off-by: NickLucche <nlucches@redhat.com>
```
  4c8f64fa
15 Jun, 2025 1 commit
- [Kernel] Raise verbose error and consolidate `num_heads/num_kv_heads` divisibility check (#19339) · 0b73736a
  22quinn authored Jun 14, 2025
```
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
```
  0b73736a
13 Jun, 2025 1 commit
- update triton_mla.py · e046b382
  zhuwenwen authored Jun 13, 2025
  
  e046b382
12 Jun, 2025 2 commits
- [torch.compile][ROCm] Fuse quantization onto attention using a torch.compile pass (#16756) · f98548b9
  Luka Govedič authored Jun 12, 2025
```
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
Co-authored-by: Sage Moore <sage@neuralmagic.com>
```
  f98548b9
- [CI] change spell checker from codespell to typos (#18711) · 2f1c19b2
  Ning Xie authored Jun 12, 2025
```
Signed-off-by: Andy Xie <andy.xning@gmail.com>
```
  2f1c19b2
11 Jun, 2025 1 commit

[AMD] [Quantization] Add override flag for attention dtype instead of using... · c7ea0b56

rasmith authored Jun 11, 2025


[AMD] [Quantization] Add override flag for attention dtype instead of using kv_cache_dtype trigger (#17331)
Signed-off-by: Randall Smith <Randall.Smith@amd.com>

c7ea0b56

06 Jun, 2025 1 commit
- [Model] Add VLLM_USE_NN to use nn layout · 6880bf15
  zhuwenwen authored Jun 06, 2025
  
  6880bf15
04 Jun, 2025 1 commit
- [CPU] V1 support for the CPU backend (#16441) · 4555143e
  Li, Jiang authored Jun 04, 2025
  
  4555143e
03 Jun, 2025 2 commits
- [V1] Support cross-layer KV sharing (#18212) · bdf13965
  Yong Hoon Shin authored Jun 03, 2025
```
Signed-off-by: Yong Hoon Shin <yhshin@meta.com>
```
  bdf13965
- [Misc] Add SPDX-FileCopyrightText (#19100) · 02f0c7b2
  Simon Mo authored Jun 03, 2025
```
Signed-off-by: simon-mo <simon.mo@hey.com>
```
  02f0c7b2
30 May, 2025 2 commits
- [ROCm] Remove unnecessary assertion of max_model_len in ROCM_AITER_MLA attention backend. (#18938) · 77b6e74f
  vllmellm authored May 30, 2025
```
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
```
  77b6e74f
- update flash_attn.py · 58042164
  zhuwenwen authored May 30, 2025
  
  58042164
29 May, 2025 2 commits
- [ROCm][V0][Attention] Revert to the previous FA triton kernel (#18226) · 1b7cfd5a
  Gregory Shtrasberg authored May 29, 2025
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
```
  1b7cfd5a
- [test]fix basic_correctness and benchmarks · dbd62f84
  zhuwenwen authored May 29, 2025
  
  dbd62f84
28 May, 2025 4 commits
- update use_rocm_custom_paged_attention · df877aad
  zhuwenwen authored May 28, 2025
  
  df877aad
- [BugFix] FA2 MLA Accuracy Issue (#18807) · ce75efee
  Lucas Wilkinson authored May 28, 2025
```
Signed-off-by: LucasWilkinson <lwilkinson@neuralmagic.com>
```
  ce75efee
- set VLLM_FLASH_ATTN_BACKEND to use FlashAttention Backend for attention computation on rocm · 82f60bef
  zhuwenwen authored May 28, 2025
  
  82f60bef
- using triton prefix cache on BW · 07b41ddf
  zhuwenwen authored May 28, 2025
  
  07b41ddf
23 May, 2025 2 commits
- tbo增加ds int8支持，修改tb zerooverhead环境变量到envs.py中统一管理 · dbbb148b
  lizhigong authored May 23, 2025
  
  dbbb148b
- 修改MLA prefill阶段出现的Device2Host拷贝同步现象。 · 3f80d9ac
  lizhigong authored May 22, 2025
  
  3f80d9ac
22 May, 2025 1 commit
- 修改MLA prefill阶段出现的Device2Host拷贝同步现象 · 0c1fa562
  zhuwenwen authored May 22, 2025
  
  0c1fa562
21 May, 2025 1 commit
- [ROCm][Kernel][V1] Enable AMD Radeon GPU Custom Paged Attention on v1 (#17004) · dd5fa7e0
  Hosang authored May 21, 2025
```
Signed-off-by: Hosang Yoon <hosang.yoon@amd.com>
```
  dd5fa7e0
20 May, 2025 1 commit
- support Z100L and K100 inference · a5aa55e8
  zhuwenwen authored May 20, 2025
  
  a5aa55e8
14 May, 2025 2 commits
- [BugFix][AMD] Compatible patch for AITER lib after 04/20 (#17912) · 4f8b3732
  qli88 authored May 14, 2025
```
Signed-off-by: Qiang Li <qiang.li2@amd.com>
```
  4f8b3732
- [Fix] Support CUDAGraph capture for encoder-decoder on ROCm (#18104) · 176a95c6
  Luka Govedič authored May 13, 2025
```
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
```
  176a95c6
13 May, 2025 2 commits
- support cutlass prefix-cache · bdc70899
  zhuwenwen authored May 13, 2025
  
  bdc70899
- Implements dual-chunk-flash-attn backend for dual chunk attention with sparse... · 60f76243
  Tao He authored May 13, 2025
```
Implements dual-chunk-flash-attn backend for dual chunk attention with sparse attention support (#11844)
```
  60f76243