Commits · bda9d0535fa7d47169c4b50c6169e59a670aeb92 · OpenDAS / vllm_cscc · GitLab

27 Jul, 2025 1 commit
- [Refactor] Refactor MOE NVFP4 Code Base: ModelOpt + Compressed Tensor (#21631) · bda9d053
  Wentao Ye authored Jul 27, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  bda9d053
26 Jul, 2025 4 commits
- [NVIDIA] Explicitly disable shuffled weights for flashinfer blockscale moe fp8 kernels (#21411) · de509ae8
  Kaixi Hou authored Jul 26, 2025
```
Signed-off-by: kaixih <kaixih@nvidia.com>
```
  de509ae8
- [Refactor] Remove `moe_align_block_size_triton` (#21335) · 56e544f2
  Wentao Ye authored Jul 26, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  56e544f2
- [Feature] Add support for MoE models in the calibration-free RTN-based quantization (#20766) · 7ae75fa6
  Alex Kogan authored Jul 25, 2025
```
Signed-off-by: Alex Kogan <alex.kogan@oracle.com>
```
  7ae75fa6
- [Perf] Cuda Kernel for Int8 Per Token Group Quant (#21476) · 75d29cf4
  Wentao Ye authored Jul 25, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  75d29cf4
25 Jul, 2025 6 commits
- [Model] Replace Mamba2 RMSNorm Gated with Fused Triton Kernel (#20839) · eab2f398
  Chih-Chieh Yang authored Jul 25, 2025
```
Signed-off-by: Chih-Chieh-Yang <7364402+cyang49@users.noreply.github.com>
Signed-off-by: Yu Chin Fabian Lim <fabian.lim@gmail.com>
Signed-off-by: Chih-Chieh Yang <7364402+cyang49@users.noreply.github.com>
Co-authored-by: Yu Chin Fabian Lim <fabian.lim@gmail.com>
```
  eab2f398
- [V1] Get supported tasks from model runner instead of model config (#21585) · 46d81d69
  Cyrus Leung authored Jul 25, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  46d81d69
- Add H20-3e fused MoE kernel tuning configs for Qwen3-Coder-480B-A35B-Instruct (#21598) · 8ed01e32
  Xu Wenqing authored Jul 25, 2025
```
Signed-off-by: 许文卿 <xwq391974@alibaba-inc.com>
```
  8ed01e32
- [Bugfix] DeepGemm utils : Fix hardcoded type-cast (#21517) · 2212cd6c
  Varun Sundar Rabindranath authored Jul 25, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  2212cd6c
- [Kernel] adding fused_moe configs for upcoming granite4 (#21332) · ce3a9b13
  Burkhard Ringlein authored Jul 25, 2025
```
Signed-off-by: Burkhard Ringlein <ngl@zurich.ibm.com>
Co-authored-by: Thomas Parnell <tpa@zurich.ibm.com>
```
  ce3a9b13
- [Bug] Fix DeepGemm Init Error (#21554) · 633f6e80
  Wentao Ye authored Jul 24, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  633f6e80
24 Jul, 2025 4 commits
- [MoE] More balanced expert sharding (#21497) · fe56180c
  Woosuk Kwon authored Jul 24, 2025
```
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
```
  fe56180c
- Update flashinfer CUTLASS MoE Kernel (#21408) · 1b25f1fe
  Shu Wang authored Jul 24, 2025
```
Signed-off-by: Shu Wang. <shuw@nvidia.com>
```
  1b25f1fe
- [Misc] Fix duplicate FusedMoEConfig debug messages (#21455) · f0f4de8f
  Nick Hill authored Jul 24, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  f0f4de8f
- [TPU][Bugfix] fix moe layer (#21340) · e74bfc70
  Chengji Yao authored Jul 24, 2025
```
Signed-off-by: Chengji Yao <chengjiyao@google.com>
Co-authored-by: Simon Mo <simon.mo@hey.com>
```
  e74bfc70
23 Jul, 2025 1 commit
- [Cleanup] Only log MoE DP setup warning if DP is enabled (#21315) · f002e9a8
  Michael Goin authored Jul 23, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  f002e9a8
22 Jul, 2025 5 commits
- [Perf] Cuda Kernel for Per Token Group Quant (#21083) · 774d0c01
  Wentao Ye authored Jul 22, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  774d0c01
- [feat]: add SM100 support for cutlass FP8 groupGEMM (#20447) · 2c8db17c
  Duncan Moss authored Jul 22, 2025
```
Signed-off-by: Duncan Moss <djm.moss@gmail.com>
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
Co-authored-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  2c8db17c
- [perf] Add fused MLA QKV + strided layernorm (#21116) · 4fb56914
  Mickaël Seznec authored Jul 22, 2025
```
Signed-off-by: Mickael Seznec <mickael@mistral.ai>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  4fb56914
- Update fp4 quantize API (#21327) · 9e23ad96
  Shu Wang authored Jul 22, 2025
```
Signed-off-by: Shu Wang <shuw@nvidia.com>
```
  9e23ad96
- Revert "[Performance] Performance improvements in non-blockwise fp8 CUTLASS MoE (#20762) (#21334) · e7b20426
  Ming Yang authored Jul 21, 2025
```
Signed-off-by: Ming Yang <minos.future@gmail.com>
```
  e7b20426
21 Jul, 2025 3 commits
- [perf] Speed up align sum kernels (#21079) · 0ec82edd
  Himanshu Jaju authored Jul 21, 2025
```
Signed-off-by: Himanshu Jaju <hj@mistral.ai>
```
  0ec82edd
- Add Nvidia ModelOpt config adaptation (#19815) · 6b46c4b6
  Zhiyu authored Jul 21, 2025
```
Signed-off-by: Zhiyu Cheng <zhiyuc@nvidia.com>
```
  6b46c4b6
- [Model][1/N] Support multiple poolers at model level (#21227) · 042af0c8
  Cyrus Leung authored Jul 21, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  042af0c8
19 Jul, 2025 5 commits
- [V1] [Hybrid] Enable piecewise CUDA Graph for mamba layers (#21194) · 881e3cbe
  Thomas Parnell authored Jul 19, 2025
```
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
```
  881e3cbe
- [NVIDIA] Add SM100 Flashinfer MoE blockscale fp8 backend for low latency (#20645) · 6d0734c5
  Kaixi Hou authored Jul 19, 2025
```
Signed-off-by: kaixih <kaixih@nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  6d0734c5
- [Kernel][Performance] Tweak MoE Batched silu_mul_fp8_quant_deep_gemm kernel (#21193) · dcc6cfb9
  Varun Sundar Rabindranath authored Jul 19, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  dcc6cfb9
- [V0 Deprecation] Remove V0 Spec Decode workers (#21152) · dd572c0a
  Woosuk Kwon authored Jul 18, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  dd572c0a
- Elastic Expert Parallel Initial Support (#20775) · 21793722
  Rui Qiao authored Jul 18, 2025
```
Signed-off-by: Rui Qiao <ruisearch42@gmail.com>
```
  21793722
18 Jul, 2025 5 commits
- [Kernel] Apply torch.Tag.needs_fixed_stride_order only for torch==2.6.0 (#19346) · b2eb2b5a
  Richard Zou authored Jul 18, 2025
```
Signed-off-by: rzou <zou3519@gmail.com>
```
  b2eb2b5a
- [Core] Set pooling params based on task and model (#21128) · 45badd05
  Cyrus Leung authored Jul 18, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  45badd05
- [Bugfix] Allocate less memory in non-batched CUTLASS MoE (#21121) · 4adc66f6
  ElizaWszola authored Jul 18, 2025
```
Signed-off-by: ElizaWszola <ewszola@redhat.com>
```
  4adc66f6
- [Core] FlashInfer CUTLASS fused MoE backend (NVFP4) (#20037) · c7d8724e
  Shu Wang authored Jul 17, 2025
```
Signed-off-by: shuw <shuw@nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  c7d8724e
- [Log] Debugging Log with more Information (#20770) · 8a8fc946
  Wentao Ye authored Jul 17, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  8a8fc946
17 Jul, 2025 6 commits
- [V0 deprecation] Remove V0 HPU backend (#21131) · 4de71463
  Woosuk Kwon authored Jul 17, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  4de71463
- [Model] Update pooling model interface (#21058) · 90bd2ab6
  Cyrus Leung authored Jul 18, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  90bd2ab6
- [Performance] Performance improvements in non-blockwise fp8 CUTLASS MoE (#20762) · 9fb2d220
  ElizaWszola authored Jul 17, 2025
```
Signed-off-by: ElizaWszola <ewszola@redhat.com>
```
  9fb2d220
- [Docs] Improve docstring formatting for `FusedMoEParallelConfig.make` (#21117) · fe8a2c54
  Harry Mellor authored Jul 17, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  fe8a2c54
- [Model] Add ToolParser and MoE Config for Hunyuan A13B (#20820) · 5a7fb3ab
  Asher authored Jul 17, 2025
```
Signed-off-by: Asher Zhang <asherszhang@tencent.com>
```
  5a7fb3ab
- [Kernel] DeepGemm MoE : Integrate triton permute / unpermute kernels (#20903) · 11dfdf21
  Varun Sundar Rabindranath authored Jul 17, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  11dfdf21