Commits · 9e27b5e4e4472d51cbac7778de86382e86f8dd48 · OpenDAS / vllm_cscc

01 Jul, 2025 1 commit
- 解决PA崩溃 · 5d3fb1d4
  zhuwenwen authored Jul 01, 2025
  
  5d3fb1d4
30 Jun, 2025 1 commit
- 解决PA 大batch，32k，崩溃问题 · 9d5187eb
  zhuwenwen authored Jun 30, 2025
  
  9d5187eb
27 Jun, 2025 1 commit
- update cutlass fa and pa · 4ddb5447
  zhuwenwen authored Jun 27, 2025
  
  4ddb5447
21 Jun, 2025 2 commits
- skip silu_and_mul_quant · d6a856f5
  zhuwenwen authored Jun 21, 2025
  
  d6a856f5
- fix build error · e87de76c
  zhuwenwen authored Jun 21, 2025
  
  e87de76c
12 Jun, 2025 2 commits
- add kvcache fp8 · 504a12b8
  zhuwenwen authored Jun 12, 2025
  
  504a12b8
- 解决大batch长seq崩溃问题 · d074a953
  zhuwenwen authored Jun 12, 2025
  
  d074a953
09 Jun, 2025 1 commit
- 增加4倍的临时空间，解决大batch 大seq时，临时空间不足访存越界的情况 · 259da693
  zhangshao authored Jun 09, 2025
  
  259da693
07 Jun, 2025 1 commit

[Kernel] Integrate CUTLASS MoE kernel with PPLX (#18762) · 84166fee

ElizaWszola authored Jun 07, 2025


Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>

84166fee

05 Jun, 2025 1 commit
- [Hardware][NVIDIA] FP4 MoE kernel optimization (#19110) · 61059bee
  Chiyue Wei authored Jun 05, 2025
```
Signed-off-by: Chiyue Wei <chiyuew@nvidia.com>
Co-authored-by: Chiyue Wei <chiyuew@nvidia.com>
```
  61059bee
04 Jun, 2025 5 commits
- [Perf] Tunings for SM100 FP8 CUTLASS kernel (#18778) · 53a5a0ce
  Michael Goin authored Jun 04, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  53a5a0ce
- Sm100 blockwise fp8 swap ab (#18564) · 5f2cd251
  Lain authored Jun 04, 2025
  
  5f2cd251
- 解决cudagraph模式下，小seq大batch PA变慢的bug · ea79ca42
  zhangshao authored Jun 04, 2025
  
  ea79ca42
- [NVIDIA] Add Cutlass MLA backend (#17625) · 41aa5784
  Kaixi Hou authored Jun 04, 2025
  
  41aa5784
- [KERNEL] Sampler. CUDA kernel for applying repetition penalty (#18437) · 5d6d1adf
  Vadim Gimpelson authored Jun 04, 2025
  
  5d6d1adf
03 Jun, 2025 3 commits
- [Perf] Tune `scaled_fp8_quant` by increasing vectorization (#18844) · e31446b6
  Michael Goin authored Jun 03, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  e31446b6
- [Kernel] DeepEP dispatch-combine kernel integration (#18434) · fa98d777
  Varun Sundar Rabindranath authored Jun 03, 2025
```
Signed-off-by: Varun <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  fa98d777
- [Misc] Add SPDX-FileCopyrightText (#19100) · 02f0c7b2
  Simon Mo authored Jun 03, 2025
```
Signed-off-by: simon-mo <simon.mo@hey.com>
```
  02f0c7b2
31 May, 2025 1 commit
- [ROCm][Kernel] Add gfx950 support for skinny gemms (#18010) · 306d6040
  Charlie Fu authored May 31, 2025
```
Signed-off-by: charlifu <charlifu@amd.com>
```
  306d6040
28 May, 2025 1 commit
- [BugFix] FA2 MLA Accuracy Issue (#18807) · ce75efee
  Lucas Wilkinson authored May 28, 2025
```
Signed-off-by: LucasWilkinson <lwilkinson@neuralmagic.com>
```
  ce75efee
27 May, 2025 3 commits
- [feat]适配sgl moe_fused_gate kernel · 0b229519
  王敏 authored May 27, 2025
  
  0b229519
- 修复0.8.5 config找不到的bug · 823b3332
  gaoqiong authored May 27, 2025
  
  823b3332
- feat(rocm-support): support mamba2 on rocm (#18565) · a547aeb8
  almersawi authored May 27, 2025
```
Signed-off-by: Islam Almersawi <islam.almersawi@openinnovation.ai>
Co-authored-by: Islam Almersawi <islam.almersawi@openinnovation.ai>
```
  a547aeb8
23 May, 2025 1 commit
- [Bugfix] Add half type support in reshape_and_cache_cpu_impl on x86 cpu platform (#18430) · d0bc2f81
  Yuqi Zhang authored May 23, 2025
```
Signed-off-by: Yuqi Zhang <yuqizhang@google.com>
Co-authored-by: Yuqi Zhang <yuqizhang@google.com>
```
  d0bc2f81
22 May, 2025 1 commit

[Build/CI] Fix CUDA 11.8 build (#17679) · 6e588da0

Tyler Michael Smith authored May 22, 2025


Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Signed-off-by: Tyler Michael Smith <tysmith@redhat.com>
Co-authored-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>

6e588da0

21 May, 2025 3 commits
- [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
- 优化pa_opt的编译速度 · 82e8ca03
  zhangshao authored May 21, 2025
  
  82e8ca03
- [Bug] Fix moe_sum signature (#18440) · 92247c52
  bnellnm authored May 21, 2025
```
Signed-off-by: Bill Nell <bnell@redhat.com>
```
  92247c52
19 May, 2025 2 commits
- fix CUDA_check redefinition in #17918 (#18287) · 258bf621
  Lucia Fang authored May 19, 2025
```
Signed-off-by: Lucia Fang <fanglu@fb.com>
Co-authored-by: Lucia (Lu) Fang <fanglu@meta.com>
```
  258bf621
- 增加PA_TC kv cache int8/fp8支持 · e02d110d
  zhangshao authored May 19, 2025
  
  e02d110d
16 May, 2025 2 commits
- [Bugfix] fix `an illegal memory access was encountered` of marlin kernel + act_order (#18245) · e73b7dfd
  Jinzhen Lin authored May 17, 2025
  
  e73b7dfd
- use ceil_div in cutlass block scaling shape check (#17918) · e23564cb
  Lain authored May 16, 2025
  
  e23564cb
15 May, 2025 1 commit

fix: typos (#18151) · a9944aab

omahs authored May 15, 2025


Signed-off-by: omahs <73983677+omahs@users.noreply.github.com>

a9944aab

14 May, 2025 5 commits
- [Kernel] Have rotary embeddings support tensors (#18046) · d93c976a
  Lucas Wilkinson authored May 14, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
```
  d93c976a
- Modularize fused experts and integrate PPLX kernels (#15956) · f9c069c8
  bnellnm authored May 14, 2025
  
  f9c069c8
- [Bugfix] fix moe marlin `topk_weight` loading (#18080) · d4154c35
  Jinzhen Lin authored May 14, 2025
```
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  d4154c35
- [AMD][torch.compile] Enable silu+fp8_quant fusion for rocm (#18082) · 7b2f28de
  Charlie Fu authored May 14, 2025
```
Signed-off-by: charlifu <charlifu@amd.com>
```
  7b2f28de
- add kvint8 · 45273722
  xiabo authored May 14, 2025
  
  45273722
13 May, 2025 2 commits
- 提升bf16 pa精度 · 98955c1f
  zhangshao authored May 13, 2025
  
  98955c1f
- Fix Broken macro for cutlass moe (#18049) · e57e4d6e
  Driss Guessous authored May 12, 2025
```
Signed-off-by: drisspg <drisspguessous@gmail.com>
```
  e57e4d6e