Commits · 63f49b8bd46b477da384c2cdd6613bf45ed3d515 · OpenDAS / vllm_cscc · GitLab

22 Mar, 2026 20 commits
- [Model Runner V2] Enable piecewise CUDA graphs for pipeline parallelism (#35162) · 63f49b8b
  zhanqiuhu authored Mar 22, 2026
```
Signed-off-by: Zhanqiu Hu <zh338@cornell.edu>
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: Woosuk Kwon <woosuk@inferact.ai>
```
  63f49b8b
- [MRV2] Use FP64 for Gumbel noise (#37798) · a5e9d511
  Woosuk Kwon authored Mar 22, 2026
```
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
```
  a5e9d511
- [Bigfix]fix lora test by pass padded size back to the layer (#37811) · c058ff44
  Yongye Zhu authored Mar 22, 2026
  
  c058ff44
- [MRV2] Skip hidden states allocation for PW CUDA graphs (#37818) · ce9b1d76
  Woosuk Kwon authored Mar 22, 2026
```
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
```
  ce9b1d76
- Enable `NemotronHPuzzle` + `NemotronHMTP` (#37803) · e74c17e1
  Netanel Haber authored Mar 22, 2026
  
  e74c17e1
- [Test] Only Run MLA model when user explicitly set for batch invariance (#37719) · eaf49786
  Wentao Ye authored Mar 22, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  eaf49786
- [Bug] Fix fp8 deepgemm batch invariant (#37718) · 77d24c4b
  Wentao Ye authored Mar 22, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  77d24c4b
- [Model Runner V2] Support multi-modal embeddings for spec decode model (#36097) · b3e84601
  Giancarlo Delfin authored Mar 22, 2026
```
Signed-off-by: Giancarlo Delfin <gdelfin@inferact.ai>
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: Woosuk Kwon <woosuk@inferact.ai>
```
  b3e84601
- [ROCm][CI] Stabilize ROCm speech-to-text translation test with lower min acc threshold (#37723) · cd1242d8
  Andreas Karatzas authored Mar 22, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  cd1242d8
- [MoE] Move PF Methods to Folder (#35927) · 4383f153
  Robert Shaw authored Mar 22, 2026
  
  4383f153
- [ROCm][CI] Make some duplicated tests optional so that they are only evaluated... · 6eedec6e
  Andreas Karatzas authored Mar 22, 2026
```
[ROCm][CI] Make some duplicated tests optional so that they are only evaluated in our nightly (#37780)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  6eedec6e
- [ROCm][CI] Added missing resampy dependency for MM audio tests (#37778) · ffc85315
  Andreas Karatzas authored Mar 22, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  ffc85315
- [ROCm][CI] get_cu_count was renamed to num_compute_units in #35042 (#37764) · 6ecba840
  Andreas Karatzas authored Mar 22, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  6ecba840
- [ROCm][CI] Fix MEGA_AOT_ARTIFACT fallback when PyTorch < 2.10.0 lacks AOT support (#37763) · 3b06c55c
  Andreas Karatzas authored Mar 22, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  3b06c55c
- [Perf] Optimize glm4.xv VIT (#37779) · b0507004
  Yang Liu authored Mar 22, 2026
```
Signed-off-by: Yang <lymailforjob@gmail.com>
```
  b0507004
- [Bugfix] Handle libsndfile sf_error(NULL) race condition in audio fallback (#37782) · 5dac719b
  Andreas Karatzas authored Mar 22, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  5dac719b
- [CI] Skip ISAAC multimodal tests due to broken upstream HF model weights (#37781) · c862481c
  Andreas Karatzas authored Mar 22, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  c862481c
- [ROCm][CI] Add large_gpu_mark to test_max_tokens_none for ROCm (#37717) · c86b17cf
  Andreas Karatzas authored Mar 21, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  c86b17cf
- [Bugfix] Fix pooling non-determinism from pinned prompt_lens aliasing (#37775) · 66f927f2
  Andreas Karatzas authored Mar 21, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  66f927f2
- [ROCm][CI] close missing quote in kernels/moe block in run-amd-test.sh (#37774) · e78bc742
  Andreas Karatzas authored Mar 21, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  e78bc742
21 Mar, 2026 17 commits
- [MoE] Move FlashInfer CuteDSL experts into fused_moe/experts/ (#37759) · 6b2fa3a7
  Robert Shaw authored Mar 21, 2026
```
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
```
  6b2fa3a7
- [Quantization][Deprecation] Remove PTPC FP8 (#32700) · eeee5b26
  Robert Shaw authored Mar 21, 2026
```
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
```
  eeee5b26
- Revert "Consolidate AWQ quantization into single awq_marlin.py file" (#37768) · 5ad04465
  Robert Shaw authored Mar 21, 2026
  
  5ad04465
- Consolidate AWQ quantization into single awq_marlin.py file · 8cc700dd
  Robert Shaw authored Mar 21, 2026
```
Merge awq.py and awq_marlin.py into a single file, eliminating the
circular import between them. awq.py becomes a backward-compat shim.
Follows the same structure as gptq_marlin.py.

Co-authored-by: Claude
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
```
  8cc700dd
- Add tensor IPC transfer mechanism for multimodal data (#32104) · 80b70884
  Brandon Pelfrey authored Mar 21, 2026
```
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <brandonpelfrey@gmail.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
```
  80b70884
- [Perf] Add SM 10.3 (B300/GB300) all-reduce communicator tuning (#37756) · 61e381dc
  Mohammad Miadh Angkad authored Mar 22, 2026
```
Signed-off-by: Mohammad Miadh Angkad <176301910+mmangkad@users.noreply.github.com>
```
  61e381dc
- [Core] Enable allreduce fusion by default for SM 10.3 (B300/GB300) (#37755) · 88f1b374
  Mohammad Miadh Angkad authored Mar 22, 2026
```
Signed-off-by: Mohammad Miadh Angkad <176301910+mmangkad@users.noreply.github.com>
```
  88f1b374
- [Hybrid] calling get_mamba_groups() once at MambaCopyBuffers.create() (#37318) · 298e5108
  Francesco Fusco authored Mar 21, 2026
```
Signed-off-by: Francesco Fusco <ffu@zurich.ibm.com>
```
  298e5108
- [ROCm] Enable DeepEP ROCm as all2allbackend for AMD GPUs. (#34692) · 3982bc2c
  Chaitanya Sri Krishna Lolla authored Mar 21, 2026
```
Signed-off-by: Tej Kiran <vpolamre@amd.com>
Co-authored-by: Tej Kiran <vpolamre@amd.com>
```
  3982bc2c
- [ROCm][CI] Update GSM8K eval config to use fp8-and-mixed models list (MI355) (#37721) · 02eec7ec
  Andreas Karatzas authored Mar 21, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  02eec7ec
- [Responses API] Add kv_transfer_params for PD disaggregation (#37424) · 17ee641c
  Bongwoo Bak authored Mar 21, 2026
```
Signed-off-by: bongwoobak <bongwoobak@gmail.com>
Co-authored-by: Chauncey <chaunceyjiang@gmail.com>
```
  17ee641c
- [ROCm][CI] Mark gemma3 as large GPU test to avoid OOM on MI250 (#37610) · 0d50fa1d
  Andreas Karatzas authored Mar 20, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  0d50fa1d
- Revert "[compile] Initialize passes at VllmBackend init" (#37733) · 1fa1e53a
  Simon Mo authored Mar 20, 2026
  
  1fa1e53a
- [ROCm][CI] Guard CudaPlatform/RocmPlatform imports to fix test collection on... · 3ffa5200
  Andreas Karatzas authored Mar 20, 2026
```
[ROCm][CI] Guard CudaPlatform/RocmPlatform imports to fix test collection on cross-platform builds (#37617)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  3ffa5200
- [MoE Refactor] Mxfp4 oracle rebased (#37128) · 87bd9189
  Yongye Zhu authored Mar 20, 2026
```
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
```
  87bd9189
- [Frontend] Remove librosa from audio dependency (#37058) · c7f98b4d
  Isotr0py authored Mar 21, 2026
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  c7f98b4d
- Add get_device_uuid for rocm (#37694) · 1c472f8f
  tmm77 authored Mar 20, 2026
```
Signed-off-by: Tiffany Mintz <Tiffany.Mintz@amd.com>
```
  1c472f8f
20 Mar, 2026 3 commits
- elastic_ep: Fix issues with repeated scale up/down cycles (#37131) · c57d38d6
  Itay Alroy authored Mar 21, 2026
```
Signed-off-by: Itay Alroy <ialroy@nvidia.com>
Co-authored-by: Ron Tourgeman <rtourgeman@nvidia.com>
```
  c57d38d6
- [BugFix] Allow qk_nope_head_dim=192 in FlashInfer MLA backend checks (#37475) · e5ed6c6c
  Kaihang Jiang authored Mar 20, 2026
```
Signed-off-by: Kaihang Jiang <kaihangj@nvidia.com>
```
  e5ed6c6c
- [Refactor] Remove unused dead code (#36171) · b3d0b379
  Wentao Ye authored Mar 20, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  b3d0b379