Compare commits

...

476 Commits

Author SHA1 Message Date
6f62c94d7e updated
Signed-off-by: Robert Shaw <rshaw@neuralmagic.com>
2025-10-03 13:47:16 -04:00
52a7d91980 debug
Signed-off-by: Robert Shaw <rshaw@neuralmagic.com>
2025-10-03 13:25:00 -04:00
d76541a6c5 Stop mergify from keeping stale PRs alive (#26169)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-10-03 16:42:34 +00:00
dd96465fd7 [BugFix][QWEN-VL]fix wrong apply_rotary_emb_torch selection introduced by #24642 (#26123)
Signed-off-by: Chendi Xue <Chendi.Xue@intel.com>
Signed-off-by: Chendi.Xue <chendi.xue@intel.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-10-03 08:52:26 -07:00
4f8f47e87e Fix undefined symbol: cutlass_moe_mm_sm100 (#26098)
Signed-off-by: Jun Jiang <jasl9187@hotmail.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-10-03 15:48:32 +00:00
d78fda7cda [Renderer] Move Processor out of LLMEngine (#26165)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-03 15:08:22 +00:00
73a99cc2a5 [Model] Fixed stream generator for gpt-oss + spec-decoding (#26027)
Signed-off-by: Aleksandr Samarin <astrlrd@nebius.com>
2025-10-03 13:43:41 +00:00
adae0c1f43 [CI/Build] do not enforce precompilation on tpu ci tests (#25992)
Signed-off-by: Xiang Si <sixiang@google.com>
2025-10-03 13:38:42 +00:00
whx
cbf9221992 [Model] Supplement to PR 24862: Pass param prefix to LLMHead (#25805)
Signed-off-by: whx-sjtu <2952154980@qq.com>
2025-10-03 21:34:53 +08:00
5f42fc53b6 [backends][short_conv] CUDA graph piecewise edits (#24215)
Signed-off-by: Paul Pak <paulpak58@gmail.com>
2025-10-03 12:59:48 +00:00
8ee846c27c [Bugfix] Re-enable prefill of max model length (#24446)
Signed-off-by: Yannick Schnider <yannick.schnider1@ibm.com>
2025-10-03 14:13:34 +02:00
812b7f54a8 [Renderer] Move Processor out of AsyncLLM (#24138)
Signed-off-by: Yang <lymailforjob@gmail.com>
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-03 11:29:45 +00:00
5f2cacdb1e Quick fix for IMA with the Prefix Prefill kernel during graph capture (#25983)
Signed-off-by: Sage Moore <sage@neuralmagic.com>
2025-10-03 11:28:22 +00:00
aa5053e3fe [Doc] Fixed shape description for fused_batched_moe.py (#25668)
Signed-off-by: Egor <e.a.krivov@gmail.com>
2025-10-03 04:00:23 -07:00
79aa244678 [Multi Modal] Configurable MM Profiling (#25631)
Signed-off-by: wwl2755 <wangwenlong2755@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-10-03 03:59:10 -07:00
kyt
2ed3f20dba [openai] Fix missing tool usage check (system message) (#24768)
Signed-off-by: kyt <eluban4532@gmail.com>
2025-10-03 18:55:44 +08:00
48f309029a [NIXL][Misc] Expose metrics from NIXL for logging to CLI (#25388)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-10-03 10:47:59 +00:00
0e93ac0b3a [CI] Fix distributed hybrid tests in CI (#26155)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
2025-10-03 09:14:18 +00:00
5446ad1d24 [test utils] correct wrong typing (#26159)
Signed-off-by: Yannick Schnider <yannick.schnider1@ibm.com>
2025-10-03 02:11:49 -07:00
f9a8084e48 [Model] Use merge_by_field_config for MM models (InternVL family) (#26153)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-03 01:59:06 -07:00
3e70e3d4d5 add(v1): RequestStatesStats to RequestOutput (#24947)
Signed-off-by: huijjj <huijong.jeong@squeezebits.com>
2025-10-03 08:56:25 +00:00
eb0fa43868 [Perf] Optimize reshape_and_cache CUDA Kernel (#25955)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Liu-congo <1502632128@qq.com>
2025-10-03 01:33:46 -07:00
0ad9951c41 [Input] Remove unused prompt field (#26097)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-03 00:23:21 -07:00
8c9117181d [Misc] Remove typing.List (#26150)
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
2025-10-03 07:00:33 +00:00
c4b48d3c0f [BUG] Reorder model config creation (#26124)
Signed-off-by: ahao-anyscale <ahao@anyscale.com>
2025-10-03 14:59:36 +08:00
10d765482d FusedMoE support for the Transformers backend (#22650)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-10-02 23:12:15 -07:00
39b643dc1a [Model] Use merge_by_field_config for MM models (G) (#26117)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-02 22:38:29 -07:00
711f485643 [Bugfix] Fix import gemm_afp4wfp4 failure on AMD (#26068)
Signed-off-by: zhewenli <zhewenli@meta.com>
2025-10-02 22:37:25 -07:00
9c5ee91b2a [ROCm] [VL] [Bugfix] Fix vit flash attn dispatcher logic for ROCm (#26104)
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
2025-10-02 22:34:53 -07:00
27edd2aeb4 [Build/CI] Revert back to Ubuntu 20.04, install python 3.12 with uv (#26103)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Simon Mo <simon.mo@hey.com>
2025-10-02 22:21:01 -07:00
e5017cd6d6 [gpt-oss] disable tool server initialization if no tool in request (#25790)
Signed-off-by: Andrew Xia <axia@meta.com>
Signed-off-by: Andrew Xia <axia@fb.com>
Co-authored-by: Andrew Xia <axia@fb.com>
2025-10-03 05:08:35 +00:00
6a7796e871 [Bug]: Limit num_reqs in dummy_run when max_num_seqs is small (#26144)
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
2025-10-03 04:00:20 +00:00
47b9339546 [DeepSeek] Improve performance of DS MLA cache kernel (#26132)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
2025-10-02 20:35:47 -07:00
5d5146eee3 [CI/Build] Conditionally register cutlass_fp4_group_mm to fix building on Hopper (#26138)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-10-02 20:32:38 -07:00
2aaa423842 [Attention] Move Backend enum into registry (#25893)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
2025-10-02 20:32:24 -07:00
ad2d788016 [Bug][Benchmark] Fix duplicate req in oversampling (#26140)
Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-10-03 02:55:24 +00:00
36ce76c632 [Log] Optimize DeepGEMM Missing Log (#26106)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-10-02 20:02:26 -06:00
f1fc2107a3 [Bugfix] Disable cascade attention with FlashInfer (#26130)
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-10-02 16:30:37 -07:00
13cdc02173 Fix MTP with deepep_low_latency (#25904)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
2025-10-02 21:29:49 +00:00
502640c3f9 [Perf] Fix and reapply move apply w8a8 block fp8 linear to class (#25696)
Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: ElizaWszola <elizaw.9289@gmail.com>
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
Signed-off-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Luka Govedič <lgovedic@redhat.com>
2025-10-02 19:35:13 +00:00
3d5f1c8640 [Mamba][KVCacheManager] Simplify kv cache manage logic for mamba + MTP (#25119)
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
2025-10-02 18:48:31 +00:00
1cab2f9cad EAGLE 3: Fix preamble so that measured speedup over Eagle 1 becomes 32% instead of 5% on MTBench (#25916)
Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
2025-10-02 11:29:35 -07:00
1e50f1be70 [Deepseek v3.2] Support indexer prefill chunking (#25999)
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
2025-10-02 10:29:12 -07:00
ad87ba927a [Small] Prevent bypassing media domain restriction via HTTP redirects (#26035)
Signed-off-by: Chenheli Hua <huachenheli@outlook.com>
2025-10-02 10:27:10 -07:00
decf7f794b [BugFix] Fix FI accuracy issue when used for MLA prefill (#26063)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
2025-10-02 17:18:13 +00:00
d00d652998 [CI/Build] Replace vllm.entrypoints.openai.api_server entrypoint with vllm serve command (#25967)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-02 10:04:57 -07:00
3b279a84be [CI] Add Blackwell DeepSeek FP8 FlashInfer MoE tests (#26040)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-10-02 09:07:19 -07:00
5e4a8223c6 [Qwen][ROCm] Flash Attention Rotary Embeddings (#24642)
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
2025-10-02 08:26:08 -07:00
e51de388a2 [Platform][CI] Added OOT platform interface e2e test that running on Ascend NPU (#25470)
Signed-off-by: leo-pony <nengjunma@outlook.com>
2025-10-02 23:19:22 +08:00
cc253b73d3 [Model] Use merge_by_field_config for MM models (D-F) (#26076)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-02 08:17:35 -07:00
7d6fb905d9 [Model] Use merge_by_field_config for MM models (A-C) (#26073)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-02 08:17:31 -07:00
418d111f8c [FA/Chore] Bump vllm-flash-attention (#25537)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-10-02 11:06:14 -04:00
be8921fbba Change size of single CUDA graph for CI to 4 (#26089)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
2025-10-02 14:14:28 +00:00
d4e7a1152d Update base image to 22.04 (jammy) (#26065)
Signed-off-by: Huy Do <huydhn@gmail.com>
2025-10-02 05:48:04 -07:00
be22bb6f3d Run:ai model streamer add GCS package support (#24909)
Signed-off-by: Peter Schuurman <psch@google.com>
2025-10-01 20:59:13 -07:00
169313b9f8 [Misc] Make handling of SamplingParams clearer in n>1 case (#26032)
Signed-off-by: Nick Hill <nhill@redhat.com>
2025-10-01 19:31:39 -07:00
0b018d8baf [ROCm][Bugfix] Add missing parameter to ROCm backend (#26029)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
2025-10-01 19:23:14 -07:00
c31246800c Support RL online quantization with torchao (#23014)
Signed-off-by: Jerry Zhang <jerryzh168@gmail.com>
2025-10-01 16:39:29 -07:00
4134312b35 [BugFix] ChunkedLocalAttention is currently not CG compatible (#26034)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-10-01 16:28:00 -07:00
da554f932e [Bug] Fix Negative Cuda Memory Usage (#25683)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-10-01 18:16:26 -04:00
aac622e0cd [ROCm][Build] Add support for AMD Ryzen AI MAX / AI 300 Series (#25908)
Signed-off-by: Hosang Yoon <hosang.yoon@amd.com>
2025-10-01 21:39:49 +00:00
1726e93ef1 [BugFix][DP/EP] Fix CUTLASS MLA hang under load (#26026)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
2025-10-01 12:30:00 -07:00
ee04c0cd04 [CI] Tweaks to GPT-OSS Eval (Blackwell) for stability (#26030)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-10-01 12:02:17 -07:00
c36f0aa300 Fix test_mamba_ssm_ssd.py due to missing _query_start_loc_to_chunk_indices_offsets (#25995)
Signed-off-by: Huamin Li <3ericli@gmail.com>
2025-10-01 18:18:36 +00:00
5234dc7451 [NVIDIA] Blackwell Family (#24673)
Signed-off-by: Johnny <johnnynuca14@gmail.com>
Signed-off-by: johnnynunez <johnnynuca14@gmail.com>
Signed-off-by: Johnny <johnnync13@gmail.com>
Signed-off-by: Salvatore Cena <cena@cenas.it>
Co-authored-by: Aidyn-A <31858918+Aidyn-A@users.noreply.github.com>
Co-authored-by: Salvatore Cena <cena@cenas.it>
2025-10-01 10:50:54 -07:00
3b7c20a6b5 [Bugfix] Apply same sampling parameters for both n=1 and n>1 (#26005)
Signed-off-by: Kenichi Maehashi <maehashi@preferred.jp>
2025-10-01 14:37:35 +00:00
f9e714813a [Benchmark] Finish documented v0.11.0 deprecation of --endpoint-type (#26007)
Signed-off-by: Nathan Scott <nathans@redhat.com>
2025-10-01 12:41:57 +00:00
2518230d3e [MISC] Fix misleading batch_size_capture_list when cuda_graph_sizes < 4 (#25829)
Signed-off-by: billishyahao <bill.he@amd.com>
Co-authored-by: Luka Govedic <ProExpertProg@users.noreply.github.com>
2025-10-01 08:39:45 -04:00
a332b84578 [CI] Only capture a single CUDA graph size in CI by default (#25951)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-10-01 10:03:44 +01:00
1405f0c7ba [Misc] Factor out common _apply_feature_select_strategy (#26003)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-01 01:31:03 -07:00
84d57342b6 [BugFix][MM] Fix Nonetype error when video is cache in qwen2.5-omni-thinker (#26004)
Signed-off-by: wwl2755 <wangwenlong2755@gmail.com>
2025-10-01 08:03:25 +00:00
57b46d769e [Doc] updating torch.compile doc link (#25989)
Signed-off-by: nadathurv <work.vnadathur@gmail.com>
Signed-off-by: WorldExplored <srreyansh.sethi@gmail.com>
Co-authored-by: Srreyansh Sethi <107075589+WorldExplored@users.noreply.github.com>
2025-10-01 07:04:56 +00:00
f48b6a03ba [Misc]allow disable pynccl (#25421)
Signed-off-by: Lu Fang <fanglu@fb.com>
Co-authored-by: Lucia (Lu) Fang <fanglu@meta.com>
2025-10-01 06:04:13 +00:00
2a69ab4899 Update to Transformers v4.56.2 (#24638)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-30 22:07:07 -07:00
8d7da92fd7 [BugFix] Fix default kv-cache-dtype default for DeepseekV3.2 (#25988)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-30 21:58:31 -07:00
e952eee698 [Bugfix] Fix __syncwarp on ROCM (#25996) 2025-09-30 21:15:11 -07:00
66bca9b8bd [MM] Add text-only mode for Qwen3-VL (#26000) 2025-09-30 21:13:42 -07:00
99028fda44 Fix INT8 quantization error on Blackwell GPUs (SM100+) (#25935)
Signed-off-by: padg9912 <phone.and.desktop@gmail.com>
2025-09-30 19:19:53 -07:00
1244948885 [Log] Optimize Log for FP8MOE (#25709)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-30 19:18:43 -07:00
a73f6491c8 Update launch_bounds_utils.h for correct compile on Multiple Cuda Arch - PTXAS out of range Warning (#25843)
Signed-off-by: Salvatore Cena <cena@cenas.it>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-30 19:18:19 -07:00
001e50c92c [Model] MTP fallback to eager for DeepSeek v32 (#25982)
Signed-off-by: Lu Fang <fanglu@fb.com>
2025-10-01 01:53:22 +00:00
96ebcaa3ad [Misc] Make EP kernels install script support uv (#25785)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-30 23:38:34 +00:00
5db1870bb9 [gpt-oss] use vLLM instead of openai types for streaming (#25186)
Signed-off-by: Andrew Xia <axia@meta.com>
Signed-off-by: Andrew Xia <axia@fb.com>
Co-authored-by: Andrew Xia <axia@fb.com>
2025-09-30 22:47:07 +00:00
2ce26b9b5d [Docs] Remove API Reference from search index (#25949)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-30 22:10:02 +00:00
a388252ac4 Add explicit pooling classes for the Transformers backend (#25322)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-30 23:07:06 +01:00
9a9f48dff7 [V1] [P/D] Add Support for KV Load Failure Recovery (#19330)
Signed-off-by: David Ben-David <davidb@pliops.com>
Co-authored-by: David Ben-David <davidb@pliops.com>
2025-09-30 14:57:08 -07:00
67f3fb0844 [Bench] Add DeepSeekV32 to MoE benchmark (#25962)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-30 14:13:48 -07:00
43b752c325 [Llama4] [multimodal] Fix misplaced dtype cast of cos_sin_cache in Llama4VisionRotaryEmbedding (#25889)
Signed-off-by: cjackal <44624812+cjackal@users.noreply.github.com>
2025-09-30 20:35:15 +00:00
cfd302db9b OffloadingConnector: Fix GPU block tracking bug (#25856)
Signed-off-by: Or Ozeri <oro@il.ibm.com>
2025-09-30 19:53:04 +00:00
fb610ae684 [Docs] Add moe kernel features doc (#25297)
Signed-off-by: Bill Nell <bnell@redhat.com>
Signed-off-by: bnellnm <49004751+bnellnm@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-30 19:03:15 +00:00
2f652e6cdf [Doc] Improve MM Pooling model documentation (#25966)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-30 18:58:29 +00:00
e6a226efba [Bug] Fix AttributeError: 'QKVParallelLinear' object has no attribute 'orig_dtype' (#25958)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-30 11:13:03 -07:00
a2e6fa7e03 [bugfix][deepseek] fix flashmla kernel selection (#25956)
Signed-off-by: youkaichao <youkaichao@gmail.com>
2025-10-01 00:30:36 +08:00
9f1c4ecaf2 [Bugfix] Token type and position embeddings fail to be applied to inputs_embeds (#25922)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-10-01 00:23:12 +08:00
ef283548f7 [Bugfix] Fix accuracy issue of TRTLLM FP8 MOE and improve logging (#25895)
Signed-off-by: Pavani Majety <pmajety@nvidia.com>
2025-09-30 10:51:31 -04:00
f4db5e6de1 [Bugfix][Model] Fix inference for Hunyuan dense models (#25354)
Signed-off-by: anion <1005128408@qq.com>
Signed-off-by: Anion <123177548+Anionex@users.noreply.github.com>
2025-09-30 14:38:07 +00:00
099aaee536 Add Hugging Face Inference Endpoints guide to Deployment docs (#25886)
Signed-off-by: sergiopaniego <sergiopaniegoblanco@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-30 14:35:06 +00:00
35fe398c7c [Kernel][Moe Configs] Add more tuned triton configs for ExpertsInt8 and FP8 (#25858)
Signed-off-by: asafg <39553475+Josephasafg@users.noreply.github.com>
2025-09-30 07:30:44 -07:00
bb6d43047e [Fix] Improve CPU backend compatibility for RISC-V (#25816)
Signed-off-by: lyd1992 <liuyudong@iscas.ac.cn>
Signed-off-by: ihb2032 <1355790728@qq.com>
2025-09-30 13:48:07 +00:00
bc546f76a1 [CI] Move applicable tests to CPU (#24080)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-30 14:45:20 +01:00
80608ba5af [NIXL] Add support for MLA caches with different latent dim (#25902)
Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
Co-authored-by: Chen Zhang <zhangch99@outlook.com>
2025-09-30 12:18:29 +00:00
e184c9c510 [perf] Use CPU tensor to reduce GPU->CPU sync (#25884)
Signed-off-by: Lehua Ding <lehuading@tencent.com>
2025-09-30 19:51:16 +08:00
d7e34b4210 [Model] Move vision_feature_select_strategy into resolve_visual_encoder_outputs (#25938)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-30 11:24:57 +00:00
ef6e0e7132 [Bugfix][Model]fix ernie45 moe gate&bias dtype to float32 (#25936)
Signed-off-by: wangyafeng <wangyafeng@baidu.com>
2025-09-30 19:11:21 +08:00
1ad3aca682 Updated TRL integration docs (#25684)
Signed-off-by: sergiopaniego <sergiopaniegoblanco@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Sergio Paniego Blanco <sergiopaniegoblanco@gmail.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-30 03:10:55 -07:00
8d0afa9b42 [Doc] Add Cambricon MLU support (#25942)
Signed-off-by: a120092009 <zhaoty0121@gmail.com>
2025-09-30 17:59:47 +08:00
fa7e254a7f [New Model] DeepSeek-V3.2 (Rebased to Main) (#25896)
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Signed-off-by: Barry Kang <43644113+Barry-Delaney@users.noreply.github.com>
Signed-off-by: Lucia Fang <fanglu@meta.com>
Co-authored-by: Chen Zhang <zhangch99@outlook.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Lucia Fang <116399278+luccafong@users.noreply.github.com>
Co-authored-by: Lucia Fang <fanglu@meta.com>
Co-authored-by: NickLucche <nlucches@redhat.com>
Co-authored-by: Siyuan Fu <siyuanf@nvidia.com>
Co-authored-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Xiaozhu Meng <mxz297@gmail.com>
Co-authored-by: Barry Kang <43644113+Barry-Delaney@users.noreply.github.com>
2025-09-30 17:14:41 +08:00
e23cacda35 [Bugfix]: Clean up chunked prefill logging when using whisper (#25075)
Signed-off-by: simondanielsson <simon.danielsson99@hotmail.com>
2025-09-30 08:17:49 +00:00
2e1b8bc2b6 [Model][Bugfix] Fix MiDashengLM audio encoder mask by removing incorrect logical_not (#25925)
Signed-off-by: zhoukz <me@zhoukz.com>
2025-09-30 08:15:23 +00:00
e47433b3c1 [BugFix] Pass config_format via try_get_generation_config (#25912) 2025-09-30 05:09:50 +00:00
23194d83e8 [BugFix] Fix DP/EP hang (#25906)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-30 04:18:59 +00:00
61aedb5ffe MoveVllmConfig from config/__init__.py to config/vllm.py (#25271)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-29 19:49:49 -07:00
d3bd171123 [Benchmark] Support benchmark throughput for external launcher DP (#25913)
Signed-off-by: Zhuohan Li <zhuohan123@gmail.com>
2025-09-30 01:43:57 +00:00
89e4050af4 [Bug] Fix Weight Loading for Block FP8 Cutlass SM90 (#25909)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-30 09:15:19 +08:00
78a47f87ce Test Prompt Embeds/LoRA compatibility and Enable LoRA Support for OPT Models (#25717)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
2025-09-30 08:10:58 +08:00
6a113d9aed [V0 Deprecation] Remove vllm.worker and update according imports (#25901) 2025-09-29 23:26:11 +00:00
2e4fe48c37 [NIXL] Increase default KV block eviction timeout on P (#25897)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-09-29 21:35:14 +00:00
8eb0a1d906 [Doc] Polish example for torchrun dp (#25899) 2025-09-29 21:31:34 +00:00
fea3e476aa [Kernel] Chunk-aligned mamba2 (#24683) 2025-09-29 23:18:25 +02:00
61a3431613 [Bugfix][ROCm] Fixing trying to import non-existent symbols from libnccl.so (#25605)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
2025-09-29 17:01:50 -04:00
9bedac9623 [Doc] Add documentation for vLLM continuous benchmarking and profiling (#25819)
Signed-off-by: Naman Lalit <nl2688@nyu.edu>
2025-09-29 20:49:49 +00:00
c42ff4f4fd [BugFix][torch.compile] KV scale calculation issues with FP8 quantization (#25513)
Signed-off-by: adabeyta <aabeyta@redhat.com>
2025-09-29 15:52:04 -04:00
d5ab28511c [Bugfix] Use correct key "ignore" for config.json non-quantized layers (#25706)
Signed-off-by: Lee Nau <lnau@nvidia.com>
2025-09-29 15:07:29 -04:00
e61eb5e09d [Model] Remove MotifForCausalLM (#25866)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-30 00:36:30 +08:00
0899ba5b42 [CI/Build] Include Transformers backend test in nightly transformers test (#25885)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-29 09:33:39 -07:00
145ac73317 [Bugfix][Speculative Decoding] Fix Eagle3 quantization config issue (#25883)
Signed-off-by: Rahul Tuli <rtuli@redhat.com>
2025-09-29 11:37:20 -04:00
d0d138bc55 [Nixl][P/D] Add cuda2cpu support (HD->DH transfer) (#24690)
Signed-off-by: Chenxi Yang <cxyang@fb.com>
Co-authored-by: Chenxi Yang <cxyang@fb.com>
2025-09-29 14:31:51 +00:00
43227236ec [torch.compile] serialize cudagraph_mode as its enum name instead of value (#25868)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
2025-09-29 13:54:52 +00:00
8616300ae2 [Model][Bugfix] Fix issues in MiDashengLM implementation for quantized models (#25854)
Signed-off-by: zhoukz <me@zhoukz.com>
2025-09-29 10:59:04 +00:00
edbaadd91f [Bugfix] Fix requirements paths in install instructions (#25827)
Signed-off-by: yingjun-mou <renzomou@gmail.com>
2025-09-29 03:49:35 -07:00
9360d34fa1 update to latest deepgemm for dsv3.2 (#25871)
Signed-off-by: youkaichao <youkaichao@gmail.com>
2025-09-29 17:51:43 +08:00
1b67b04656 [Misc] Remove more get_input_embeddings_v0 (#25857)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-29 08:03:37 +00:00
bd51f78e39 [V0 Deprecation][Models] Remove all V0 condition for mm embeddings merge (#25331)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: isotr0py <2037008807@qq.com>
2025-09-29 14:09:18 +08:00
65ecb4f134 [Bugfix] Fallback ViT attn backend to SDPA for blackwell (#25851)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-29 06:03:51 +00:00
143844fa43 [XPU]Fix xpu spec decoding UTs, avoid using cuda graph (#25847)
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
2025-09-29 05:15:10 +00:00
219cfbe7f6 Add Phi4FlashForCausalLM to _PREVIOUSLY_SUPPORTED_MODELS (#25832)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
2025-09-29 05:08:17 +00:00
9b44a7d926 [P/D] NIXL Updates (#25844)
Signed-off-by: Sage Moore <sage@neuralmagic.com>
Signed-off-by: simon-mo <simon.mo@hey.com>
Signed-off-by: rentianyue-jk <rentianyue-jk@360shuke.com>
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Chenheli Hua <huachenheli@outlook.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Sage Moore <sage@neuralmagic.com>
Co-authored-by: Russell Bryant <rbryant@redhat.com>
Co-authored-by: rentianyue-jk <rentianyue-jk@360shuke.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Chenheli Hua <huachenheli@outlook.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
2025-09-29 04:46:30 +00:00
a3ae45a38c [Misc] fix tests failure by using current_platform (#25825)
Signed-off-by: Juechen Liu <jueliu@meta.com>
2025-09-29 04:18:57 +00:00
0307428d65 Remove redundant cudagraph dispatcher warning (#25841) 2025-09-28 17:12:42 -04:00
471997adf6 [Bugfix] fix Qwen3VLMoe load when pp > 1 (#25838)
Signed-off-by: liuye.hj <liuye.hj@alibaba-inc.com>
Co-authored-by: liuye.hj <liuye.hj@alibaba-inc.com>
2025-09-28 17:56:12 +00:00
b1ded114b9 Update GLM-4.5 Doc transformers version (#25830)
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
2025-09-28 12:05:51 +00:00
f4e4088c99 Fix random dataset mismatched token length with config. (#24937)
Signed-off-by: Weiliang Liu <weiliangl@nvidia.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-28 08:23:44 +00:00
0efd540dbc [VLM] Update Qwen3-VL max_num_video_tokens calculation for configurable video profiling (#25557)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-28 04:21:01 +00:00
6144754014 [Bugfix] Fix Qwen3-VL regression from #24982 (#25814)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-28 03:21:09 +00:00
69311446ba [MM] Optimize memory profiling for scattered multimodal embeddings (#25810)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-28 02:17:58 +00:00
da63274d9f [Bugfix][NIXL] Fix Async Scheduler timeout issue (#25808)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-09-27 15:17:35 -04:00
c216119d64 [Core] GC Debug callback (#24829)
Signed-off-by: Jialin Ouyang <jialino@meta.com>
Signed-off-by: Jialin Ouyang <Jialin.Ouyang@gmail.com>
Co-authored-by: Jialin Ouyang <jialino@meta.com>
2025-09-27 17:53:31 +00:00
5546acb463 [Bug]: Set LD_LIBRARY_PATH to include the 'standard' CUDA location (#25766)
Signed-off-by: Clayton Coleman <smarterclayton@gmail.com>
2025-09-27 13:36:28 -04:00
c0ec81836f [torch.compile]: Add VLLM_DEBUG_DUMP_PATH environment variable (#25651)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Signed-off-by: Jiangyun Zhu <riverclouds.zhu@qq.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-27 16:09:00 +00:00
b65e56babe [Core] Refactor self.model() to call a helper for subclassing. (#25084)
Signed-off-by: Patrick Toulme <ptoulme@meta.com>
Signed-off-by: Patrick Toulme <pctoulme+1@gmail.com>
2025-09-27 08:40:59 -07:00
49996cd597 [env] default nixl side port conflicts with kv-event zmq port (#25056)
Signed-off-by: Peter Pan <Peter.Pan@daocloud.io>
2025-09-27 15:02:40 +00:00
ecb37e276a [docs] transcriptions API audio upload (#25446)
Signed-off-by: zxw <1020938856@qq.com>
2025-09-27 15:00:35 +00:00
a5354b3ed2 [Bugfix][WideEP] Apply TP Attn + EP MoE fix to other models (#24982)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2025-09-27 14:22:28 +00:00
f9df8b4ad7 [Bugfix] Fix triton import precommit failure (#25803)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2025-09-27 07:13:11 -07:00
ec152c8748 Fix GPTQ model loading in Transformers backend (#25770)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-27 12:18:20 +00:00
7977e5027c Add filtering for chat template kwargs (#25794)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-27 10:46:49 +00:00
3f5d902d2a Validate API tokens in constant time (#25781)
Signed-off-by: rentianyue-jk <rentianyue-jk@360shuke.com>
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Co-authored-by: rentianyue-jk <rentianyue-jk@360shuke.com>
2025-09-27 18:09:26 +08:00
27d7638b94 [Bugfix] Merge MM embeddings by index instead of token IDs (#16229)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: NickLucche <nlucches@redhat.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-27 08:15:12 +00:00
176173989a [Bugfix] Add missing image_size for phi4_multimodal (#25796) 2025-09-27 07:59:22 +00:00
23b8ee672d [Misc] Update openai client example file for multimodal (#25795)
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-27 07:57:07 +00:00
3939152069 [Misc] Fix codeowners override for v1 sample and attention (#25037)
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
2025-09-27 07:47:29 +00:00
cd87bfbf37 [CI/Build] Reorganize root-level V1 tests (#25767)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-27 13:51:15 +08:00
b3613e3ace [CI/Build] Add timing to Model Executor Test (#25799)
Signed-off-by: 22quinn <33176974+22quinn@users.noreply.github.com>
2025-09-26 21:57:27 -07:00
d346ec695e [CI/Build] Consolidate model loader tests and requirements (#25765)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-26 21:45:20 -07:00
c242c98031 [Bugfix] Allow Only SDPA Backend for ViT on B200 for Qwen3-VL (#25788) 2025-09-26 20:44:52 -07:00
f1d53d150c [Multimodal][Speculative Decoding]Eagle Eagle3 mm support, enablement on qwen2.5vl (#22872)
Signed-off-by: Junhong <liujunhong11@huawei.com>
Signed-off-by: Junhong Liu <98734602+LJH-LBJ@users.noreply.github.com>
Co-authored-by: Junhong <liujunhong11@huawei.com>
Co-authored-by: LJH-LBJ <98734602+LJH-LBJ@users.noreply.github.com>
2025-09-27 03:35:47 +00:00
92da847cf5 Add flashinfer-build.sh and register precompiled cu128 wheel in Dockerfile (#25782)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-26 18:54:09 -07:00
3958b96bf5 Add option to restrict media domains (#25783)
Signed-off-by: Chenheli Hua <huachenheli@outlook.com>
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Co-authored-by: Chenheli Hua <huachenheli@outlook.com>
2025-09-27 01:23:52 +00:00
8bf8f45822 [Core] Don't count preempted tokens in prefix cache hit rate (#25787)
Signed-off-by: Zhuohan Li <zhuohan123@gmail.com>
2025-09-27 00:16:40 +00:00
6f5c0931c1 [Spec decode] automatically disable mm for text-only draft models (#25667)
Signed-off-by: Jonas Kuebler <kuebj@amazon.com>
2025-09-27 08:10:21 +08:00
4e33a7ea85 [Bugfix] Optimize CpuGpuBuffer initialization (#25447)
Signed-off-by: Naman Lalit <nl2688@nyu.edu>
2025-09-27 08:07:36 +08:00
dc48ba0c75 Kernel-override Determinism [1/n] (#25603)
Signed-off-by: Bram Wasti <bwasti@meta.com>
2025-09-26 16:59:09 -07:00
4778b42660 Reduce the Cuda Graph memory footprint when running with DBO (#25779)
Signed-off-by: Sage Moore <sage@neuralmagic.com>
2025-09-26 22:29:56 +00:00
c70ac4b8ff [spec decode] Consolidate speculative decode method name for MTP (#25232)
Signed-off-by: zixi-qi <qizixi@meta.com>
2025-09-26 22:27:05 +00:00
cf89202855 [CI] Fix FlashInfer AOT in release docker image (#25730)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-26 14:11:40 -07:00
f075693da7 [V1] address post issues related to #20059 (part 1) (#23046)
Signed-off-by: fhl2000 <63384265+fhl2000@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-26 15:58:19 -04:00
f708bd4904 [CI] Add E2E Blackwell Quantized MoE Test (#25723)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-26 12:23:00 -07:00
0002b7f0d1 [Docs] Add Toronto Meetup (#25773)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-26 12:00:46 -07:00
11aafd9886 [Bugfix] Improve GLM4 MoE Reasoning Parser's is_reasoning_end Condition (#25355)
Signed-off-by: frankwang28 <frank.wbb@hotmail.com>
Signed-off-by: Frank Wang <41319051+frankwang28@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Chauncey <chaunceyjiang@gmail.com>
2025-09-26 11:54:00 -07:00
b761df963c [Doc]: improve CPU(x86) build-wheel-from-source section (#25617)
Signed-off-by: Kosseila (CloudThrill) <klouddude@gmail.com>
2025-09-26 10:26:33 -07:00
33f6aaf972 Eagle3 that supports the Minicpm3 model (#24243)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: liudan <adan@minicpm.com>
Co-authored-by: liudan <liudan@qq.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Lucia Fang <116399278+luccafong@users.noreply.github.com>
2025-09-26 10:04:57 -07:00
56aafa8c0b [Misc] fix unique_filepath (#25732)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-26 16:56:15 +00:00
8d52f2b3a7 [ray][metrics] Replace ':' with '_' for OpenTelemetry compatibility in Ray (#25439)
Signed-off-by: Seiji Eicher <seiji@anyscale.com>
Signed-off-by: Seiji Eicher <58963096+eicherseiji@users.noreply.github.com>
Co-authored-by: Rui Qiao <161574667+ruisearch42@users.noreply.github.com>
2025-09-26 09:43:30 -07:00
984d18498a [BugFix] Fix using dbo_decode_token_threshold always (and ignoring dbo_prefill_token_threshold) (#25622)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-26 16:22:49 +00:00
d4d9899860 [Quantization] Add field to skip unquantized modules for GPTQ config (#25455)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-26 15:47:41 +00:00
db1e42f627 [CI/Build] Fix some V1 tests not being run (#25569)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-26 20:52:36 +08:00
bc9d7b5595 [CI/Build] Split up Distributed Tests (#25572)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-26 14:49:33 +02:00
fe6b19c314 [Bugfix] Properly abort pooling request. (#25734)
Signed-off-by: wang.yuqi <noooop@126.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
2025-09-26 05:47:34 -07:00
2827b3f4a3 [CI] Fix test_shared_storage_connector_hashes (#25748)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
2025-09-26 20:46:17 +08:00
2b6b1d7809 [Model] Mamba2 varlen refactor (#21467)
Signed-off-by: Chih-Chieh-Yang <7364402+cyang49@users.noreply.github.com>
Co-authored-by: RishiAstra <40644327+RishiAstra@users.noreply.github.com>
2025-09-26 11:31:14 +00:00
633f943e30 [Doc] Update Batch-level DP docs (#25757)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-26 02:37:40 -07:00
b03b1b97f6 Support LongCat-Flash-Chat tool call (#24083)
Signed-off-by: 许文卿 <xwq391974@alibaba-inc.com>
2025-09-26 09:25:39 +00:00
dfb9af2014 [Bugfix] Fix Shared Expert/Zero expert code in FusedMoE.process_chunk (#25698)
Signed-off-by: Sage Moore <sage@neuralmagic.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
2025-09-26 01:25:28 -07:00
19f76ee68e [misc] refactor speculative config (#25657)
Signed-off-by: zxw <1020938856@qq.com>
2025-09-26 01:22:06 -07:00
dd70437a4f Remove cuda hard-code in compute_causal_conv1d_metadata (#25555)
Signed-off-by: Icey <1790571317@qq.com>
2025-09-26 01:19:20 -07:00
99b3a504c5 [Qwen3-Next][GDN] fixes cuda graph capturing bug in GDN metadata and a stride bug in causal_conv_1d. (#25743)
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
2025-09-26 01:18:58 -07:00
6e30010d2f fix: print outputt offline_inference/base/chat.py example (#25744)
Signed-off-by: Iceber Gu <caiwei95@hotmail.com>
2025-09-26 01:18:24 -07:00
52621c8f5c [Harware][AMD][Model] Triton MoE tuning configs for GLM-4.5 for MI300X (#25703)
Signed-off-by: xaguilar <Xavier.AguilarFruto@amd.com>
2025-09-26 01:18:20 -07:00
d48f4d6daf perf: Avoid copying inputs_embeds tensors to GPU unless prompt_embeds is enabled (#25739)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
2025-09-26 01:18:09 -07:00
e84e0735c7 fix: revert cast to cpu in MsgpackEncoder._encode_tensor to avoid hidden performance regressions (#25738)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
2025-09-26 01:18:05 -07:00
3edf87d25f [CI/Build] fix doc build warning: Failed to get 'name: description' pair (#25733)
Signed-off-by: yiting.jiang <yiting.jiang@daocloud.io>
2025-09-26 01:18:02 -07:00
392edee34a EVS Support (Video tokens pruning) (#22980)
Signed-off-by: Eugene Khvedchenia <ekhvedchenia@nvidia.com>
Signed-off-by: Eugene Khvedchenya <ekhvedchenya@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-26 11:54:54 +08:00
983056e456 [Misc] Remove unnecessary memoryviews in shm_broadcast.py (#25721)
Signed-off-by: Nick Hill <nhill@redhat.com>
2025-09-26 03:11:44 +00:00
13dd93c667 [Core] Force PIECEWISE CUDAGraph mode for encoder-decoder (#25701)
Signed-off-by: Russell Bryant <rbryant@redhat.com>
2025-09-25 18:21:56 -07:00
53a30845be Llamas 3.1 405B fp4 changes upstreaming from 355_wip (#25135)
Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Doug Lehr <douglehr@amd.com>
2025-09-25 19:16:53 -06:00
8b77328ffe [Misc] Don't log shm dequeue delay warning on worker side (#25720)
Signed-off-by: Nick Hill <nhill@redhat.com>
2025-09-26 01:08:30 +00:00
9fe4c2bdb9 [Refactor] Remove DeepGEMM OP Register (#25710)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-25 20:13:41 -04:00
081b5594a2 Fix routing_bias dtype (#25711)
Signed-off-by: Shu Wang. <shuw@nvidia.com>
2025-09-25 23:35:14 +00:00
57329a8c01 [Model] rename NemotronH_Nano_VL -> NemotronH_Nano_VL_V2 (#25708)
Signed-off-by: Tomer Asida <57313761+tomeras91@users.noreply.github.com>
2025-09-25 16:10:29 -07:00
8c435c9bce [Core] Enable command line logging for LLMEngine (#25610)
Co-authored-by: Ye (Charlotte) Qi <yeq@meta.com>
Signed-off-by: Zhuohan Li <zhuohan123@gmail.com>
2025-09-25 15:31:17 -07:00
e71b8e210d [Spec Decode] Add Batch Parallel Ngram. Upto 8x lower overhead. (#24986)
Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
Co-authored-by: Nick Hill <nhill@redhat.com>
2025-09-25 15:22:03 -07:00
89fa54e6f7 [Optimization] Use a cheaper cache key in get_model_architecture (#25682)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 17:54:20 -04:00
3d54bdcb73 [Optimization] Streamline InputPreprocessor (#25702)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 21:06:49 +00:00
6b0fcbbf43 [Misc] Simplify test_argsort_mm_positions (#25690)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 18:23:01 +00:00
0fa673af4c [V0 deprecation] Clean up LoRA (#25686)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-25 18:12:33 +00:00
3468f17ebe [V0 deprecation] Remove _VLLM_V1 suffixes from attention backend names (#25489)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Signed-off-by: Matthew Bonanni <mbonanni001@gmail.com>
2025-09-25 17:37:50 +00:00
71b25b0d48 [V0 deprecation] Clean up V0 fallback in compilation config (#25675)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-25 17:29:51 +00:00
0ea80c87d9 [Model] Define merge_by_field_config MM interface (#25676)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 17:13:07 +00:00
b8d9e4a326 [Model] Add optional parameter to reasoning parser constructor (#25554)
Signed-off-by: taohui <taohui3@gmail.com>
Signed-off-by: Tao Hui <taohui3@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-26 01:12:50 +08:00
13cc7f5370 [BugFix] Fix DBO hang (#25625)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-25 17:04:48 +00:00
916bd9204d Revert "[Bug] Dynamo Unsupported due to BasevLLMParameter.torch_function calling disabled super()" (#25681)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
2025-09-25 09:45:06 -07:00
e04a1b6b21 [BUGFIX] Fix crash in Eagle Speculative Decoding models when exceedin… (#24662)
Signed-off-by: AlonKejzman <alonkeizman@gmail.com>
2025-09-25 15:40:14 +00:00
2e5df88c92 [Logging] Remove TORCH_NCCL_AVOID_RECORD_STREAMS to squash a warning (#25532)
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
2025-09-25 15:16:06 +00:00
0754ac4c49 [Misc] Remove cruft file in repo (#25678)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-09-25 08:05:12 -07:00
03858e6d1c [Bugfix] Fix InternS1 video processing after Transformers v4.56 (#25644)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-25 14:46:04 +00:00
532a6cfccb [ux] Switch a warning to debug about a pytorch fallback (#23750)
Signed-off-by: Russell Bryant <rbryant@redhat.com>
2025-09-25 14:38:16 +00:00
eb32335e35 [CPU] update torch 2.8 and fix missing fields in TorchSDPAMetadata (#25652)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
2025-09-25 13:29:11 +00:00
69a8c8e99a [torch.compile] Make Query Quantization Fusable (#24914)
Signed-off-by: Jonas Kuebler <kuebj@amazon.com>
2025-09-25 09:25:12 -04:00
6c340da4df [misc] log info messages by default for hanging / busy / idle (#25627)
Signed-off-by: youkaichao <youkaichao@gmail.com>
2025-09-25 21:14:57 +08:00
2f17117606 [mypy] Fix wrong type annotations related to tuple (#25660)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 13:00:45 +00:00
1e9a77e037 [Hardware][RISC-V] Add riscv64 support for vLLM with scalar (#22112)
Signed-off-by: chenlang <chen.lang5@zte.com.cn>
Co-authored-by: chenlang <10346245@zte.com.cn>
2025-09-25 20:46:11 +08:00
d2af67441d [XPU][Triton]add xpu config in triton_reshape_and_cache_flash (#25643)
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
2025-09-25 12:38:11 +00:00
0bcc3a160d [CI/Build] Fix flaky entrypoints test (#25663)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 12:19:40 +00:00
70fbdb26e9 Add backward compatibility for guided_... API (#25615)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
2025-09-25 19:45:25 +08:00
7f570f1caa [V0 deprecation] Remove unreachable model_config.supported_tasks (#25642)
Signed-off-by: wang.yuqi <noooop@126.com>
2025-09-25 11:26:31 +00:00
eaeca3cd7f [Bugfix] Parse SpeculativeConfig Error (#25142)
Signed-off-by: zxw <1020938856@qq.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-25 11:09:39 +00:00
12c1287d64 [mypy] Further improve MM type annotations (#25654)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 10:57:36 +00:00
17b4c6685c [Bugfix] Fix Qwen3-VL max_num_video_tokens calculation for video profiling (#25648)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-25 18:36:01 +08:00
3c2b2ccece [Bugfix] Add triton.language.tensor placeholder (#25649)
Signed-off-by: Agata Dobrzyniewicz <adobrzyniewicz@habana.ai>
2025-09-25 10:31:14 +00:00
7be9ffcd9f [Misc] Fix Qwen3-VL video_grid_thw typing (#25646)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-25 10:16:45 +00:00
393de22d2e [fix] Update torch version in cpu-build.txt for AArch64/ppc64le and Darwin (#25579)
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
2025-09-25 09:39:18 +00:00
1260180c67 Revert "[Performance] Move apply_w8a8_block_fp8_linear to an op class… (#25607)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2025-09-25 08:05:21 +00:00
af4ee63e0e typo: remove duplicate is (#25641)
Signed-off-by: nicole-lihui <nicole.li@daocloud.io>
2025-09-25 00:46:22 -07:00
bc092ea873 Map CwmForCausalLM to llama and LlamaForCausalLM (#25611)
Signed-off-by: Jacob Kahn <jacobkahn1@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-25 07:37:03 +00:00
755ed7b05b [Misc] Simplify PoolerOutput and move to v1/outputs (#25629)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-25 06:47:03 +00:00
a676e668ee [Bugfix] fix apply_temperature to avoid nan in probs (#24734)
Signed-off-by: courage17340 <courage17340@163.com>
2025-09-25 05:32:21 +00:00
c85be1f6dd optimize: eliminate duplicate split_enc_dec_inputs calls (#25573)
Signed-off-by: nicole-lihui <nicole.li@daocloud.io>
2025-09-25 05:03:25 +00:00
845adb3ec6 [Model] Add LongCat-Flash (#23991)
Signed-off-by: yangxurui <yangxurui@meituan.com>
Co-authored-by: yangxurui <yangxurui@meituan.com>
2025-09-24 21:53:40 -07:00
90b139cfff Enable Fbgemm NVFP4 on Dense models (#25609)
Signed-off-by: Saman Keon <samanamp@outlook.com>
2025-09-24 21:12:53 -07:00
4492e3a554 [Bug] Dynamo Unsupported due to BasevLLMParameter.torch_function calling disabled super() (#25613)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-24 18:52:52 -07:00
05c19485a5 [Kernel] Support DCP for Triton backend (#25132)
Signed-off-by: Wei Wei <wwei6@meta.com>
2025-09-24 18:09:34 -07:00
52d0cb8458 [Model] Improve DotsOCRForCausalLM (#25466)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-25 07:58:08 +08:00
5c1e496a75 [MISC] replace c10::optional with std::optional (#25602)
Signed-off-by: Shiyan Deng <dsy842974287@meta.com>
2025-09-24 16:56:21 -07:00
e7f27ea648 Improve --help for enhanced user experience (#24903)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-24 23:08:18 +00:00
1f29141258 [Refactor] Use DeepGEMM Col Major TMA Aligned Tensor (#25517)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-24 18:52:36 -04:00
6160ba4151 feat: BF16 FlashInfer Fused Cutlass MOE for Hopper and Blackwell Expert Parallel (#25503)
Signed-off-by: Duncan Moss <djm.moss@gmail.com>
2025-09-24 18:50:04 -04:00
fea8006062 [Logging] Improve log for when DeepEP HT disables CUDA Graphs (#25531)
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
2025-09-24 22:43:06 +00:00
e6750d0b18 [V0 Deprecation] Remove unused classes in attention (#25541)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
2025-09-24 13:24:40 -07:00
8c853050e7 [Docs] Enable fail_on_warning for the docs build in CI (#25580)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-24 19:30:33 +00:00
f84a472a03 Suppress benign cuBLAS warning when capturing cudagraphs with DBO (#25596)
Signed-off-by: Sage Moore <sage@neuralmagic.com>
2025-09-24 19:02:08 +00:00
54e42b72db Support mnnvl all2allv from Flashinfer (#21003)
Signed-off-by: Shu Wang <shuw@nvidia.com>
Signed-off-by: Shu Wang. <shuw@nvidia.com>
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2025-09-24 14:38:16 -04:00
2dda3e35d0 [Bugfix] add cache model when from object storage get model (#24764)
Signed-off-by: rongfu.leng <rongfu.leng@daocloud.io>
2025-09-24 18:11:16 +00:00
d83f3f7cb3 Fixes and updates to bench_per_token_quant_fp8 (#25591)
Signed-off-by: Michael Goin <mgoin64@gmail.com>
2025-09-24 08:30:15 -07:00
302eb941f3 [ROCm][Build][Bugfix] Fix ROCm base docker whls installation order (#25415)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
2025-09-24 11:25:10 -04:00
487745ff49 [ROCm][Bugfix] Only enable +rms_norm based on aiter if not explicitly disabled (#25275)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
2025-09-24 11:24:39 -04:00
9313be5017 [Misc] Improve type annotations for jsontree (#25577)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-24 22:49:58 +08:00
8938774c79 Move DeviceConfig, ObservabilityConfig, SpeechToTextConfig to their own files (#25564)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-24 13:59:05 +00:00
e18b714b2e [Bugfix] Fix DeepSeekV31ToolParser to correctly parse multiple tools in non-streaming output (#25405)
Signed-off-by: taohui <taohui3@gmail.com>
2025-09-24 20:58:00 +08:00
b1068903fd [docs] fix nixl kv_connector_extra_config.backends key (#25565)
Signed-off-by: Peter Pan <Peter.Pan@daocloud.io>
Signed-off-by: Peter Pan <peter.pan@daocloud.io>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-24 11:00:27 +00:00
164299500b [Benchmark] Fix regression in structured output benchmark (#25500)
Signed-off-by: Russell Bryant <rbryant@redhat.com>
2025-09-24 10:40:42 +00:00
58c360d9be [Bug] fix import and unit test (#25558)
Signed-off-by: Jonas M. Kübler <44084297+jmkuebler@users.noreply.github.com>
2025-09-24 10:17:59 +00:00
42488dae69 [Bugfix] Fix dummy video number of frames calculation (#25553)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-24 09:47:30 +00:00
b67dece2d8 [misc] update the warning message (#25566)
Signed-off-by: youkaichao <youkaichao@gmail.com>
2025-09-24 17:24:35 +08:00
2338daffd3 [BugFix] Potential Fix for FA3 full-cudagraph IMA (#25490)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-24 02:04:04 -07:00
2e19a848d4 [V0 Deprecation] Remove max_seq_len_to_capture (#25543)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-24 01:51:39 -07:00
77a7fce1bb [CI/Build] add nightly prime-rl integration tests (#25207)
Signed-off-by: Jackmin801 <ongjackm@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-24 08:44:22 +00:00
6488f3481b [Misc]] Move processing context to multimodal directory (#25548)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-24 08:15:00 +00:00
27ec3c78f3 [CI/Build] Fix v1 OOT registration test (#25547)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-24 08:03:13 +00:00
1cbcfb94de [Bugfix][CPU] Skip unsupported custom op register on CPU (#25534)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
2025-09-24 06:21:51 +00:00
fed8a9b107 [Misc] Retry HF processing if "Already borrowed" error occurs (#25535)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-23 22:32:11 -07:00
190c45a6af [TPU][Bugfix] fix the missing apply_model in tpu worker (#25526)
Signed-off-by: Chengji Yao <chengjiyao@google.com>
2025-09-24 05:18:08 +00:00
5caaeb714c [Bugfix] [Frontend] Cleanup gpt-oss non-streaming chat tool calls (#25514)
Signed-off-by: Ben Browning <bbrownin@redhat.com>
2025-09-24 03:20:38 +00:00
d747c2ef18 [Perf] Fix jit compiles at runtime of fla gated delta rule (#25432)
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-24 11:16:13 +08:00
c30b405b8f [Spec Decode] Enable FlashInfer Spec Decoding (#25196)
Signed-off-by: Benjamin Chislett <benjamin.chislett@centml.ai>
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
Co-authored-by: lhsjohn <huashuoli@tencent.com>
2025-09-23 22:29:58 -04:00
77d906995c [KV sharing] Re-land Gemma3n model changes from #22628 (#24357)
Signed-off-by: Yong Hoon Shin <yhshin@meta.com>
2025-09-23 19:25:34 -07:00
359d293006 [fix]: add Arm 4bit fused moe support (#23809)
Signed-off-by: Nikhil Gupta <nikhil.gupta2@arm.com>
2025-09-24 01:32:22 +00:00
9df8da548e [BugFix] Fix MLA assert with CUTLASS MLA (#25478)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-23 21:09:43 -04:00
bf68fd76a9 [Compile] Fix AMD Compile Error (#25518)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-24 00:42:48 +00:00
de94289a98 [Core] Support weight_loader_v2 for UnquantizedLinearMethod (#23036)
Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
2025-09-23 18:30:26 -06:00
1983609239 [Bugfix] Use a separate FlashInfer workspace buffer for trtllm-gen (#25520) 2025-09-24 00:19:56 +00:00
d06b5a95cb [V1][Metrics] Add per-request TPOT histogram (#24015)
Signed-off-by: baxingpiaochong <771405853@qq.com>
2025-09-23 18:19:04 -06:00
be0bb568c9 [Model] Support SeedOss Reason Parser (#24263)
Signed-off-by: Yan Lu <luyan@nvidia.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-23 18:15:51 -06:00
c8bde93367 [BUG] Allows for RunAI Streamer and Torch.compile cache to be used together (#24922)
Signed-off-by: ahao-anyscale <ahao@anyscale.com>
2025-09-23 18:13:32 -06:00
88d7bdbd23 [Bug] Fix AttributeError: 'FusedMoE' object has no attribute 'w13_weight_scale'. Did you mean: 'w13_weight_scale_inv' (#25519)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-24 00:07:51 +00:00
0d235b874a Add CUTLASS FP8 MOE benchmark scripts and kernel config (#25302)
Signed-off-by: Chenxi Yang <cxyang@fb.com>
Co-authored-by: Chenxi Yang <cxyang@fb.com>
2025-09-23 18:07:42 -06:00
7ad5e50adf Improve output when failing json.loads() on structured output test (#25483)
Signed-off-by: dougbtv <dosmith@redhat.com>
2025-09-23 18:03:31 -06:00
dc464a3d39 [BugFix] AssertionError: Do not capture num_reqs > max_num_reqs for uniform batch (#25505)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-23 18:00:29 -06:00
1210e4d95b [Bugfix] [B200] cutlass_mla - ensure kv_split == 1 for batch size > 1 (#25509)
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
2025-09-23 16:57:55 -07:00
e0b24ea030 [Perf] Increase default max splits for FA3 full cudagraphs (#25495)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
2025-09-23 16:53:34 -07:00
bde2a1a8a4 [ROCm] Small functional changes for gptoss (#25201)
Signed-off-by: jpvillam <jpvillam@amd.com>
Co-authored-by: jpvillam <jpvillam@amd.com>
2025-09-23 23:39:50 +00:00
5e25b12236 [Kernel] [Mamba] Remove BLOCK_H=1 from list of tuneable configurations for _chunk_cumsum_fwd_kernel (#25197)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
Co-authored-by: Chih-Chieh-Yang <chih.chieh.yang@ibm.com>
2025-09-23 23:23:30 +00:00
c85d75cf08 Add VLLM_NVTX_SCOPES_FOR_PROFILING=1 to enable nvtx.annotate scopes (#25501)
Signed-off-by: Corey Lowman <clowman1993@gmail.com>
2025-09-23 22:50:09 +00:00
abad204be6 [BugFix] Fix OOM in vLLM replicas by ensuring consistent NCCL memory accounting (#25359)
Signed-off-by: Kourosh Hakhamaneshi <kourosh@anyscale.com>
2025-09-23 15:49:09 -07:00
7361ab379f Remove redundant mutates_args and dispatch_key for direct_register_custom_op (#25512)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-23 22:48:40 +00:00
95bc60e4cb [gpt-oss][bugfix] remove logic to require resp_ in ResponseAPI (#25428)
Signed-off-by: Andrew Xia <axia@meta.com>
2025-09-23 15:46:46 -07:00
4f2954f724 Fix triton_reshape_and_cache_flash.py triton import (#25522)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-23 15:26:10 -07:00
eca7be9077 Add VLLM_ENABLE_INDUCTOR_MAX_AUTOTUNE & VLLM_ENABLE_INDUCTOR_COORDINA… (#25493)
Signed-off-by: rouchenzi <ruochenwen@gmail.com>
Signed-off-by: rouchenzi <40842833+rouchenzi@users.noreply.github.com>
2025-09-23 22:17:49 +00:00
969b4da3a6 [V0 Deprecation] Remove placeholder attn (#25510)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
2025-09-23 22:12:14 +00:00
4f8c4b890a [Core] Use KVCacheBlock as much as possible instead of dict[block_id, KVCacheBlock] (#24830)
Signed-off-by: Jialin Ouyang <Jialin.Ouyang@gmail.com>
2025-09-23 15:11:14 -07:00
ae002924e9 [CI/Build] Fix and re-enable v1 PP test on CI (#25496)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-23 21:58:25 +00:00
690f948e4a [Bugfix] Fix for the import error from #24588 (#25481)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
2025-09-23 21:31:08 +00:00
08275ec0a2 [Build] Update Xgrammar to 0.1.25 (#25467)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
2025-09-23 21:25:46 +00:00
c828d1bf98 [Bugfix] gpt-oss container tool output bug (#25485)
Signed-off-by: Alec Solder <alecs@fb.com>
Co-authored-by: Alec Solder <alecs@fb.com>
2025-09-23 20:43:45 +00:00
8b8a8afc89 [CI] Fix Pre-commit Issue (#25497)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-24 04:09:37 +08:00
8bdd8b5c51 Enable symmetric memory all reduce by default only enabling for TP (#25070)
Signed-off-by: ilmarkov <markovilya197@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-23 15:53:00 -04:00
a8ffc4f0f2 [Bugfix] Lower gpt-oss max cudagraph size to 992 to be compatible with FA3 (#25508)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-23 12:49:55 -07:00
d5944d5146 [Speculators][Speculative Decoding] Fix gpt-oss eagle3 accuracy issue (#25406)
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
2025-09-23 15:44:35 -04:00
24fab45d96 [Perf] Change default CUDAGraphMode from PIECEWISE to FULL_AND_PIECEWISE (#25444)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-23 15:29:26 -04:00
63400259d0 [Performance] Move apply_w8a8_block_fp8_linear to an op class (#24666)
Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: ElizaWszola <elizaw.9289@gmail.com>
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
Signed-off-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Luka Govedič <lgovedic@redhat.com>
2025-09-23 12:03:10 -07:00
8c1c81a3de [core] add nccl symmetric memory for all reduce (#24532)
Signed-off-by: Amir Samani <asamani@nvidia.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-23 14:33:06 -04:00
a3a7828010 [ROCm] Add skinny gemm bias support for dtypes fp16,bf16,fp8 (#24988)
Signed-off-by: Hashem Hashemi <hashem.hashemi@amd.com>
Signed-off-by: Hashem Hashemi <159079214+amd-hhashemi@users.noreply.github.com>
2025-09-23 14:31:45 -04:00
5abb117901 [Core] Ensure LoRA linear respect the base_layer's tp_size and tp_rank (#25487)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-23 18:19:25 +00:00
867ecdd1c8 [Spec Decode][CI] Add e2e test for examples/spec_decode.py and prevent breaking Acceptance Length (#24531)
Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-23 10:46:40 -07:00
24e8222745 [Misc] Reduce initialization time of auto_tune (#23682)
Signed-off-by: Weida Hong <wdhongtw@google.com>
2025-09-23 17:34:58 +00:00
100b630a60 [V1][Kernel] Add triton implementation for reshape_and_cache_flash (#24503)
Signed-off-by: Burkhard Ringlein <ngl@zurich.ibm.com>
Co-authored-by: Chih-Chieh Yang <chih.chieh.yang@ibm.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
2025-09-23 12:52:40 -04:00
527821d191 Use macro guard CUDA functions for back compatibility in grouped_topk_kernel.cu (#25346)
Signed-off-by: Ming Yang <minos.future@gmail.com>
Signed-off-by: Rahul Tuli <rtuli@redhat.com>
Co-authored-by: Rahul Tuli <rtuli@redhat.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Lu Fang <30275821+houseroad@users.noreply.github.com>
Co-authored-by: Ye (Charlotte) Qi <yeq@meta.com>
2025-09-23 09:45:39 -07:00
846197f505 [Log] Optimize kv cache memory log from Bytes to GiB (#25204)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-23 12:44:37 -04:00
2357480b1a [BugFix] Fix UB in per_token_group_quant.cu (#24913)
Signed-off-by: Shreeasish Kumar <shreeasish@rivosinc.com>
2025-09-23 09:14:22 -07:00
f11e3c516b [Kernels] Support blocked fp8 quantization for compressed tensors MoE (#25219)
Signed-off-by: Bill Nell <bnell@redhat.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-23 16:11:34 +00:00
875d6def90 Add backward compatibility for GuidedDecodingParams (#25422)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-23 17:07:30 +01:00
cc1dc7ed6d [Core/DBO][2/N] Dual-Batch Overlap add DeepEP High Throughput support and Prefill support (#24845)
Signed-off-by: Sage Moore <sage@neuralmagic.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Sage Moore <sage@neuralmagic.com>
Co-authored-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
2025-09-23 16:02:10 +00:00
a903669e10 [V1] Remove V0 code paths for Hybrid models (#25400)
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
2025-09-23 08:26:13 -07:00
2c58742dff [UX] Change kv-cache-memory log level to debug (#25479)
Signed-off-by: Michael Goin <mgoin64@gmail.com>
2025-09-23 08:01:24 -07:00
4c966e440e [XPU] Fix MOE DP accuracy issue on XPU (#25465) 2025-09-23 14:32:57 +00:00
da5e7e4329 [Docs] NixlConnector quickstart guide (#24249)
Signed-off-by: Peter Pan <Peter.Pan@daocloud.io>
Signed-off-by: Peter Pan <peter.pan@daocloud.io>
Signed-off-by: Nicolò Lucchesi<nicolo.lucchesi@gmail.com>
Co-authored-by: Nicolò Lucchesi <nicolo.lucchesi@gmail.com>
2025-09-23 14:23:22 +00:00
f05a4f0e34 [P/D] Support NIXL connector to disconnect during a clean shutdown (#24423)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Co-authored-by: Mark McLoughlin <markmc@redhat.com>
2025-09-23 16:08:02 +02:00
61d1b35561 [BugFix] Register expert_map as named buffer for wake_up and sleep (#25458)
Signed-off-by: wuxibin <wuxibin@bytedance.com>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
2025-09-23 21:49:13 +08:00
b6a136b58c [CI/Build] Fix disabled v1 attention backend selection test (#25471)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-23 13:05:46 +00:00
0d9fe260dd [docs] Benchmark Serving Incorrect Arg (#25474)
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
2025-09-23 06:05:11 -07:00
273690a50a [Core] Optimize LoRA weight loading (#25403)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-23 18:19:45 +08:00
231c2c63e4 [Bugfix] Fix idefics3 tie_word_embeddings (#25454)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-23 10:06:48 +00:00
4322c553a6 [Test]: Hermes tool parser stream output error in Qwen3 case (#25203)
Signed-off-by: Andreas Hartel <andreas.hartel@aleph-alpha.com>
2025-09-23 17:56:31 +08:00
babad6e5dd [Misc] Move DP for ViT code inside model executor dir (#25459)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-23 09:20:52 +00:00
9383cd6f10 [Frontend] Add a new xml-based tool parser for qwen3-coder (#25028)
Signed-off-by: Zhikaiiii <1658973216@qq.com>
2025-09-23 16:07:27 +08:00
ba8d2165b6 Handle triton kernel import exception (#25319)
Signed-off-by: Ming Yang <minos.future@gmail.com>
2025-09-23 00:56:00 -07:00
c98be0a232 [Model] Enable DP for ViT in Qwen2-VL (#25445)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-23 05:17:10 +00:00
5774b0a1da [NIXL][OOT platform] support nixl_connector with oot platform and other nixl_backend (#25121)
Signed-off-by: Chendi Xue <Chendi.Xue@intel.com>
2025-09-23 04:17:42 +00:00
e8db44f883 [DP/EP][GPTOSS] Use triton matmul-ogs kernels for GPTOSS DP/EP (#24588)
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
2025-09-22 21:01:09 -07:00
fafbe11af4 [Docs] Fix griffe warnings in vllm/lora/ops (#25369)
Signed-off-by: windsonsea <haifeng.yao@daocloud.io>
2025-09-23 03:42:58 +00:00
78237e43bf [Bugfix] Remove contiguous output req for context parallel MLA (#25414)
Signed-off-by: Michael Goin <mgoin64@gmail.com>
2025-09-22 20:26:32 -07:00
eea1783989 [benchmarks]allow skip ready check for bench serve (#25420)
Signed-off-by: Lu Fang <fanglu@fb.com>
Signed-off-by: Lucia Fang <116399278+luccafong@users.noreply.github.com>
Co-authored-by: Lucia (Lu) Fang <fanglu@meta.com>
2025-09-23 03:21:48 +00:00
f225ea7dd9 [XPU] Fix compile_size is None case. (#25433)
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
2025-09-23 03:09:00 +00:00
fc97733da8 [feat] Support MRoPE + YaRN (#25384)
Signed-off-by: liuye.hj <liuye.hj@alibaba-inc.com>
Co-authored-by: liuye.hj <liuye.hj@alibaba-inc.com>
2025-09-23 03:04:47 +00:00
4741239db7 [Bug] Fix Long Context OOM Issue (#25290)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-22 22:04:15 -04:00
c625f9043c [V0 deprecation] Remove _set_default_args_v0 function (#25409)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-23 01:52:09 +00:00
6fa78d8f23 [V0 deprecation] Remove platform v1 controling interface (#25410)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-23 01:48:12 +00:00
9949aa2ef1 [Perf] Apply torch.compile for per_block_cast_to_fp8 (#24611)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-22 19:42:45 -06:00
0b7bed9c38 [Performance] Remove input pads in cutlass_mla and optimize v_proj output handling (#25184)
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
2025-09-22 19:20:53 -06:00
ac0048c0ae [BugFix] [DP/EP] Fix slow execution when BS <= DP (#25407)
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Chris Bamford <chrisbam4d@gmail.com>
2025-09-22 17:26:17 -07:00
090197034f [Bugfix] Fix missing clear_connector_metadata (#25397)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-09-23 08:10:59 +08:00
f31ff87460 [Core] Drop overly aggressive whisper assertion (#25408)
Signed-off-by: Russell Bryant <rbryant@redhat.com>
2025-09-22 17:09:52 -07:00
d588cd2406 [Bugfix] fix custom op test (#25429)
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
2025-09-23 00:07:43 +00:00
45d7d852d3 [Frontend] Responses API MCP tools for built in tools and to pass through headers (#24628)
Signed-off-by: Alec Solder <alecs@fb.com>
Signed-off-by: Alec S <10566873+alecsolder@users.noreply.github.com>
Co-authored-by: Alec Solder <alecs@fb.com>
Co-authored-by: Ye (Charlotte) Qi <yeq@meta.com>
2025-09-22 23:38:19 +00:00
8bed179109 [TPU] update torch_xla dependency for PyPI compatibility (#25278)
Signed-off-by: Johnny Yang <johnnyyang@google.com>
Co-authored-by: Chengji Yao <chengjiyao@google.com>
2025-09-22 16:14:44 -07:00
f552d5e578 [CI/Build] Skip Qwen3-VL initialization tests until models are actually released (#25394)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-22 13:18:24 -07:00
8db2939289 [KV offload][5/N] Add CPUOffloadingSpec (#24251)
Signed-off-by: Or Ozeri <oro@il.ibm.com>
2025-09-22 12:30:36 -07:00
d5e0fca264 [torch.compile] Cleanup compilation tests and custom passes, add debug utils, fix DCE bug (#23091), fix test (#24376), and prep for custom op matching (#24604) (#24542)
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
Signed-off-by: luka <lgovedic@redhat.com>
Signed-off-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-22 12:30:05 -07:00
8d0ee5a564 [misc] Remove RFC review hours reference (#25416) 2025-09-22 12:16:59 -07:00
922979bfcc [DP] support torchrun external launcher with Data Parallelism (#24899)
Signed-off-by: Lu Fang <fanglu@fb.com>
Signed-off-by: Zhuohan Li <zhuohan123@gmail.com>
Co-authored-by: Zhuohan Li <zhuohan123@gmail.com>
2025-09-22 12:06:05 -07:00
239ef0c1ac [CI Failure] Fix fp8 kv cache on <SM90 (#25396)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-22 18:27:51 +00:00
1d7f95b85c [Compiler] Disable Inductor standalone compile by default (#25391)
Signed-off-by: ElizaWszola <ewszola@redhat.com>
2025-09-22 17:37:46 +00:00
cfbee3d0e7 [CLI env var] Add VLLM_FLASH_ATTN_MAX_NUM_SPLITS_FOR_CUDA_GRAPH in env variables (#25274)
Signed-off-by: qqma <qqma@amazon.com>
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Co-authored-by: qqma <qqma@amazon.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
2025-09-22 10:37:43 -07:00
06a41334c7 [EPLB] Reduce EPLB Inference Overhead (#24573)
Signed-off-by: Bowen Wang <abmfy@icloud.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
2025-09-22 16:31:05 +00:00
175811e3b5 [V1][Attention] Split triton_attn in triton-only and rocm specific backends (#24648)
Signed-off-by: Burkhard Ringlein <ngl@zurich.ibm.com>
2025-09-22 15:20:28 +00:00
c10101a3eb [Bugfix] Fix several issues with p2p xPyD in GET type (#23993)
Signed-off-by: Csrayz <jover@cmbchina.com>
Signed-off-by: ivyilike <pww123@cmbchina.com>
Co-authored-by: ivyilike <pww123@cmbchina.com>
2025-09-22 14:53:13 +00:00
ac243886b0 [Kernel] MI-300X triton moe configs (#23445)
Signed-off-by: Sara Kokkila Schumacher <saraks@ibm.com>
2025-09-22 14:29:54 +00:00
3d2c56b7a9 Make mypy behave like a proper pre-commit hook (#25313)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-22 12:23:45 +00:00
64c824cd78 Make pickle import check fast (#25379)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-22 04:08:25 -07:00
417a164af6 [Misc] Remove unused encoder-decoder error strings (#25374)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-22 11:04:32 +00:00
b6f01bd9a7 refactor: abstract graph mode support into platform interface (#25161)
Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
2025-09-22 10:22:29 +00:00
4cf71cc88a [TPU] Deprecate xm.mark_step in favor of `torch_xla.sync (#25254)
Signed-off-by: NickLucche <nlucches@redhat.com>
Co-authored-by: Ye (Charlotte) Qi <yeq@meta.com>
2025-09-22 10:12:57 +00:00
a66d131381 [TPU][Bugfix][CI] Fix broken tests/build dependency (#25255)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-09-22 09:55:04 +00:00
21467f9a1c Enable Eagle3 speculative decoding for GPT-OSS model (#25246)
Signed-off-by: Eldar Kurtic <8884008+eldarkurtic@users.noreply.github.com>
2025-09-22 08:50:39 +00:00
f92d952632 [V0 Deprecation] Remove MultiModalPlaceholderMap (#25366)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-22 08:49:19 +00:00
6d0b827cbd [V0 Deprecation] Remove V0-only methods in multi-modal registry (#25362)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-22 13:58:26 +08:00
0eecb31663 [Bugfix] Fix hermes tool parser handling of non-string argument types (#22002)
Signed-off-by: wangzi <3220100013@zju.edu.cn>
Signed-off-by: David Chen <530634352@qq.com>
Co-authored-by: wangzi <3220100013@zju.edu.cn>
Co-authored-by: Chauncey <chaunceyjiang@gmail.com>
2025-09-22 11:35:39 +08:00
793be8d057 [Docs] GSM8K Accuracy Evaluation doc update (#25360)
Signed-off-by: David Chen <530634352@qq.com>
2025-09-22 02:49:13 +00:00
7b57a433da [Model] Support Dots OCR (#24645)
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: yinz-aizip <yinz@aizip.ai>
2025-09-22 02:24:40 +00:00
5aeb925452 Multimodal - audio tests (#25285)
Signed-off-by: Debolina Roy <debroy@redhat.com>
2025-09-22 07:07:11 +08:00
04d3752329 [Bugfix][V0 Deprecation][CI] use async mock and await for async method (#25325)
Signed-off-by: Yang <lymailforjob@gmail.com>
2025-09-22 07:06:16 +08:00
bc6e542d9f Remove V0 attention backends (#25351)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-21 16:03:28 -07:00
af7dfb0d1a [Perf] Further optimization for Qwen3-VL fast_pos_embed_interpolate (#25347)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-21 20:12:45 +00:00
1c3ffdbecc [V0 Deprecation] Remove V0 sampling metadata (#25345)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
2025-09-21 10:37:11 -07:00
c438b2951c feat: Enable engine-level arguments with speculators models (#25250)
Signed-off-by: Rahul Tuli <rtuli@redhat.com>
Co-authored-by: Claude <noreply@anthropic.com>
2025-09-21 11:04:45 -06:00
0ff8ebb2d7 [V0 Deprecation] Remove async_output_proc, preemption mode, delay factor (#25334)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-21 08:52:32 -07:00
26e673fe93 [V0 Deprecation] Remove V0 Sequence class & Sampler (#25332)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
2025-09-21 08:52:15 -07:00
65a5910ce3 [Optimization] Cache chat template result when processor fails to be loaded (#25341)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-21 19:41:02 +08:00
9aea7373ff [Bugfix] Typos in error message for missing model config file (#25339)
Signed-off-by: simondanielsson <simon.danielsson99@hotmail.com>
2025-09-21 04:36:47 -07:00
30d08911f7 [MM][Perf] Minor Optimization on Qwen3-VL fast_pos_embed_interpolate (#25337)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-21 11:05:20 +00:00
cf56cf78b4 [V1] Add sliding window support to Flex Attention backend (#24089)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-21 05:08:07 +00:00
7ed82d1974 [V0 Deprecation] Remove V0 MP executor (#25329)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 21:26:35 -07:00
12dbd834cf [V0 Deprecation] Remove from_seq_group methods (#25330)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 21:10:48 -07:00
035fd2bd2c [Multi Modal][Performance] Fused Q,K's apply_rope in more models (#25005)
Signed-off-by: wwl2755 <wangwenlong2755@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-21 03:55:10 +00:00
1cd885bd54 [V0 Deprecation] Remove V0 model runner base & simplify worker base (#25328)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 20:49:09 -07:00
62b38dc832 [Doc] improve test-pipeline.yaml documentation (#25305)
Signed-off-by: Huamin Li <3ericli@gmail.com>
Co-authored-by: Lu Fang <30275821+houseroad@users.noreply.github.com>
2025-09-20 20:29:12 -07:00
c99db8c8dd [V0 Deprecation] Remove V0 core (#25321)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 19:58:26 -07:00
72dd1595b4 [CI] Skip tests failing on main (#25326)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 19:57:46 -07:00
572ddf83ce [Chore] Remove unused sampler in models (#25324)
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 19:53:20 -07:00
86647d1cd0 [V0 Deprecation] Remove V0 Output Processor (#25320)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 17:57:20 -07:00
52c2a8d4ad [V0 Deprecation] Remove LLMEngine (#25033)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
2025-09-20 17:56:30 -07:00
367a480bd3 [Docs] Fix warnings in vllm/profiler and vllm/transformers_utils (#25220)
Signed-off-by: windsonsea <haifeng.yao@daocloud.io>
2025-09-20 16:39:47 -07:00
bef180f009 [V0 Deprecation] Enable the remaining multimodal tests in V1 (#25307)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-20 17:50:58 +00:00
d88918e4c2 [Core] Enable sharded state loader for V1 engine and enhance test coverage (#25308)
Signed-off-by: pengdrumli <pengdrumli@tencent.com>
2025-09-20 21:15:22 +08:00
3c713a9711 [Model] Cleanup InternViT's data parallel implementation (#25306)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-20 05:46:24 -07:00
bf8b26cad1 Generate _ModelInfo properties file when loading to improve loading speed (#23558)
Signed-off-by: Manoel Marques <manoel.marques@ibm.com>
Signed-off-by: Manoel Marques <manoelmrqs@gmail.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-20 11:51:13 +00:00
032d661d27 [Docs] Fix warnings in mkdocs build (continued) (#25042)
Signed-off-by: wwl2755 <wangwenlong2755@gmail.com>
2025-09-20 11:45:18 +00:00
e08a3a3fdb [CI Failure] Disable FlashInfer RoPE to unblock CI (#25299)
Signed-off-by: mgoin <mgoin64@gmail.com>
2025-09-20 08:16:56 +00:00
3d9a1d2de5 [V1] Support LLM.apply_model (#18465)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-20 07:14:35 +00:00
be874c0201 [Bugfix] Fix Qwen3-VL-MoE weight loading for EP (#25300)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-20 00:04:05 -07:00
9607d5eb44 [Hybrid Allocator] Support full attention with different hidden size (#25101)
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
2025-09-19 23:43:59 -07:00
c60e6137f0 [Optimization] Avoid repeated model architecture conversion for pooling models (#25261)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-20 13:30:22 +08:00
f91480b2d4 [Bugfix] fix tool call arguments is empty (#25223)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Co-authored-by: xin.li <xin.li@daocloud.io>
2025-09-20 13:29:54 +08:00
6c5f82e5aa [BUG FIX][NON-CUDA]quick fix to avoid call cudagraph_unsafe in attention (#25298)
Signed-off-by: Chendi Xue <Chendi.Xue@intel.com>
2025-09-20 04:41:23 +00:00
b7f186bbb3 [BugFix] Exclude self when checking for port collision (#25286)
Signed-off-by: Nick Hill <nhill@redhat.com>
2025-09-20 12:28:31 +08:00
3642909617 [BUGFIX] GPTQ quantization compatibility for Qwen3 Next MOE models (AutoGPTQ and AutoRound-GPTQ) (#25268)
Signed-off-by: JartX <sagformas@epdcenter.es>
2025-09-20 11:18:13 +08:00
c308501cb6 Improve weight loading for encoder models in Transformers backend (#25289)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-20 03:11:03 +00:00
535d80056b [Misc] Support more collective_rpc return types (#25294)
Signed-off-by: Nick Hill <nhill@redhat.com>
2025-09-20 02:02:38 +00:00
a25ade5d47 [BugFix] Ensure appropriate guards in destructors (#25284)
Signed-off-by: Nick Hill <nhill@redhat.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-20 09:06:34 +08:00
8945b001db [torch.compile] CUDAGraph Inductor partition integration (#24281)
Signed-off-by: Boyuan Feng <boyuan@meta.com>
Signed-off-by: Boyuan Feng <fby.1994@gmail.com>
Signed-off-by: boyuanfeng <boyuan@meta.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-20 01:02:15 +00:00
b8a287a0a8 [docs] Prompt Embedding feature support (#25288)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
2025-09-19 17:46:23 -07:00
c7e713616a test: Remove vestigial skip for prompt embeds tests after landing v1 Prompt Embeds support (#25291)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
2025-09-19 17:33:40 -07:00
a36c675817 Don't skip special tokens with hermes-style tool calling (#25281)
Signed-off-by: Max de Bayser <mbayser@br.ibm.com>
2025-09-19 17:33:25 -07:00
3da17c2cc2 [Bugfix] Remove VLLM_TEST_DYNAMO_FULLGRAPH_CAPTURE #2969 (#25090)
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
2025-09-19 20:27:21 -04:00
14c1432789 [BugFix] Fix async scheduling CPU tensor race take 2 (#25279)
Signed-off-by: Nick Hill <nhill@redhat.com>
2025-09-19 16:34:07 -07:00
ee7a66dd9a allow disable flashinfer prefill (#25276)
Signed-off-by: Lu Fang <fanglu@fb.com>
2025-09-19 22:59:41 +00:00
431535b522 Enable modelopt gemma3 nvfp4/fp8, make workflow more robust (#22771)
Signed-off-by: Zhiyu Cheng <zhiyuc@nvidia.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-19 22:40:33 +00:00
711e912946 [Compile] Fix Compile Warning for Ignoring MIN_BLOCK_PER_SM (#25193)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2025-09-19 16:23:19 -06:00
e69e0b8b5f [Frontend] Responses API messages out, just harmony for now (#24985)
Signed-off-by: Alec Solder <alecs@fb.com>
Co-authored-by: Alec Solder <alecs@fb.com>
Co-authored-by: Ye (Charlotte) Qi <yeq@meta.com>
2025-09-19 21:40:16 +00:00
ddc9048394 Fix: Correct FusedMoE layer reference in auto_round quantization (#24818)
Signed-off-by: David-Wen <18927700430@163.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-09-19 20:44:24 +00:00
b1a63d1b3b [BugFix] Make FlashInferMetadataBuilder non-blocking (#25040)
Signed-off-by: Julien Lin <jullin@nvidia.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
2025-09-19 20:36:34 +00:00
48ecb4438b [Perf] Use FlashInfer RoPE for RotaryEmbedding.forward_cuda when available (#21126)
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
2025-09-19 14:06:49 -06:00
e57fc15971 Specify platform in pip-compile pre-commit hook so it runs on MacOS (#25273)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-19 12:43:33 -07:00
4bdf400218 [Bugfix] Fix chunked a2_scales in modular kernels (#25264)
Signed-off-by: Bill Nell <bnell@redhat.com>
2025-09-19 19:42:01 +00:00
7852b82b93 [Bugfix] GPT OSS Attritbute error on H100 (#25228)
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
2025-09-19 13:14:09 -06:00
a2a5f79e09 Optimize triton unified attention performance for sliding window attention (#24390)
Signed-off-by: zixi-qi <qizixi@meta.com>
2025-09-19 13:07:26 -06:00
c59a0eca42 [KV offload][4/N] Offloading KV connector (#22595)
Signed-off-by: Or Ozeri <oro@il.ibm.com>
2025-09-19 19:07:17 +00:00
b716ab93a7 [bugfix] fix structured outputs key missing issue from #24929 (#25195)
Signed-off-by: Lu Fang <fanglu@fb.com>
2025-09-19 18:37:57 +00:00
138f0d1e75 [Docs] add __init__.py to vllm/model_executor/layers/quantization/compressed_tensors/transform (#24974)
Signed-off-by: samzong <samzong.lu@gmail.com>
2025-09-19 18:32:27 +00:00
2506ce5189 [Core][Prefix Hash] Fix prefix hash metrics sliding window maintainance (#24990)
Signed-off-by: Jialin Ouyang <Jialin.Ouyang@gmail.com>
2025-09-19 12:22:53 -06:00
47fd08aaf9 [CI/Build] fix test function_calling (#25072)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
2025-09-19 12:16:32 -06:00
12aed7e453 Encoder model support for the Transformers backend (#25174)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-19 19:15:22 +01:00
d90e212a3a Remove Redundant Assignment in Qwen3_VisionPatchMerger (#25224)
Signed-off-by: Junhong <liujunhong11@huawei.com>
Co-authored-by: Junhong <liujunhong11@huawei.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
2025-09-19 12:15:13 -06:00
2821986450 [Core] Modify the initialization parameters of the lora manager (#25249)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
2025-09-19 18:01:28 +00:00
6c117cff7d [Frontend] Pass API server count to each process (#23717)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-20 01:15:19 +08:00
7ac67ea525 [KV offload][3/N] Add worker-side CPU support (#21448)
Signed-off-by: Or Ozeri <oro@il.ibm.com>
2025-09-19 09:53:45 -07:00
ce75e15373 refactor(benchmarks): add type annotations to wait_for_endpoint parameters (#25218)
Signed-off-by: samzong <samzong.lu@gmail.com>
2025-09-19 16:36:52 +00:00
aed16879a9 Move ModelConfig from config/__init__.py to config/model.py (#25252)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-19 16:22:33 +00:00
cf278ff3b2 Update CODEOWNERS (#25269)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-19 09:12:55 -07:00
838d7116ba [Qwen] Remove cuda hard-code in qwen3 next (#25243)
Signed-off-by: Icey <1790571317@qq.com>
2025-09-19 12:25:12 +00:00
5089fd749c [V0 Deprecation] Remove V0 logic from get_input_embeddings interface (#25242)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-19 11:10:52 +00:00
a3d087adec [P/D][Nixl] Introduce KVTransferMetrics and aggregation strategy (#22188)
Signed-off-by: NickLucche <nlucches@redhat.com>
2025-09-19 11:09:14 +00:00
058525b997 Move PoolerConfig from config/__init__.py to config/pooler.py (#25181)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
2025-09-19 11:02:55 +00:00
1dfea5f4a9 [Bugfix][Perf] Misc fixes for Qwen3 VL (#25238)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-19 10:46:16 +00:00
cea91a32f2 [Kernel][Performance] Add Triton kernel for Qwen3-VL interleaved MRoPE (#25055)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-19 10:27:49 +00:00
a684c0124c [bugfix] fix MHA for models like OpenGVLab/InternVL3_5-38B (#25146)
Signed-off-by: Yan Ma <yan.ma@intel.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-19 08:45:06 +00:00
f2718d2948 [Misc] Cleanup test conftest for deprecated encoder-decoder models (#25231)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
2025-09-19 07:44:56 +00:00
825fdb11ad [Bugfix][CPU] Add placeholder to avoid import errors when using fused_moe ops on platforms without triton (#25137)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
2025-09-19 07:41:12 +00:00
8c1d4acbfe [CPU] Disable oneDNN linear on non-x86 platforms (#25166)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
2025-09-19 07:27:22 +00:00
486c5599e3 [Build] Update Xgrammar to 0.1.24 to get a CVE fix (#25188)
Signed-off-by: Russell Bryant <rbryant@redhat.com>
2025-09-19 14:27:17 +08:00
a6149aa587 [OOT] Support sync_model_loading for OOT (#25126)
Signed-off-by: Chendi Xue <Chendi.Xue@intel.com>
2025-09-19 05:41:53 +00:00
6c8a3c099b [Docs] Fix griffe warnings in vllm/multimodal (#25216)
Signed-off-by: windsonsea <haifeng.yao@daocloud.io>
2025-09-18 22:10:44 -07:00
31a8a2a7bc [Misc] Clean up MM profiling warnings (#25222)
Signed-off-by: Roger Wang <hey@rogerw.io>
2025-09-19 04:46:57 +00:00
1a0a04dae9 [Perf] Optimize memory peak during EAGLE model loading. (#24585)
Signed-off-by: Chen Ding <candy.dc@alibaba-inc.com>
2025-09-19 03:31:16 +00:00
6d8246aaff [gpt-oss] Add ResponseReasoningPartAddedEvent, ResponseReasoningPartDoneEvent for streaming (#24938)
Signed-off-by: Andrew Xia <axia@meta.com>
2025-09-18 19:11:59 -07:00
9d1c50a5ac [KV offload][2/N] Introduce LRU-based CPU offloading management (#20075)
Signed-off-by: Or Ozeri <oro@il.ibm.com>
2025-09-19 00:20:51 +00:00
9a4600e4dc [CORE] Prompt Embeddings Support for v1 Engine (#24278)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
Signed-off-by: Andrew Sansom <qthequartermasterman@gmail.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
2025-09-19 08:03:09 +08:00
992 changed files with 49006 additions and 47334 deletions

View File

@ -181,18 +181,14 @@ launch_vllm_server() {
if echo "$common_params" | jq -e 'has("fp8")' >/dev/null; then
echo "Key 'fp8' exists in common params. Use neuralmagic fp8 model for convenience."
model=$(echo "$common_params" | jq -r '.neuralmagic_quantized_model')
server_command="python3 \
-m vllm.entrypoints.openai.api_server \
server_command="vllm serve $model \
-tp $tp \
--model $model \
--port $port \
$server_args"
else
echo "Key 'fp8' does not exist in common params."
server_command="python3 \
-m vllm.entrypoints.openai.api_server \
server_command="vllm serve $model \
-tp $tp \
--model $model \
--port $port \
$server_args"
fi

View File

@ -365,8 +365,7 @@ run_serving_tests() {
continue
fi
server_command="$server_envs python3 \
-m vllm.entrypoints.openai.api_server \
server_command="$server_envs vllm serve \
$server_args"
# run the server

View File

@ -76,7 +76,7 @@ steps:
queue: arm64_cpu_queue_postmerge
commands:
- "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list='8.7 9.0 10.0+PTX 12.0' --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
- "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg FLASHINFER_AOT_COMPILE=true --build-arg torch_cuda_arch_list='8.7 9.0 10.0+PTX 12.0' --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
- "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)"
# Add job to create multi-arch manifest

View File

@ -86,10 +86,6 @@ if [[ $commands == *"pytest -v -s models/test_registry.py"* ]]; then
commands=${commands//"pytest -v -s models/test_registry.py"/"pytest -v -s models/test_registry.py -k 'not BambaForCausalLM and not GritLM and not Mamba2ForCausalLM and not Zamba2ForCausalLM'"}
fi
if [[ $commands == *"VLLM_USE_V1=0 pytest -v -s models/test_initialization.py -k 'not llama4 and not plamo2'"* ]]; then
commands=${commands//"VLLM_USE_V1=0 pytest -v -s models/test_initialization.py -k 'not llama4 and not plamo2'"/"VLLM_USE_V1=0 pytest -v -s models/test_initialization.py -k 'not llama4 and not plamo2 and not BambaForCausalLM and not Gemma2ForCausalLM and not Grok1ModelForCausalLM and not Zamba2ForCausalLM and not Gemma2Model and not GritLM'"}
fi
if [[ $commands == *"pytest -v -s compile/test_basic_correctness.py"* ]]; then
commands=${commands//"pytest -v -s compile/test_basic_correctness.py"/"VLLM_USE_TRITON_FLASH_ATTN=0 pytest -v -s compile/test_basic_correctness.py"}
fi

View File

@ -58,11 +58,8 @@ function cpu_tests() {
# pytest -x -v -s tests/kernels/attention/test_cache.py -m cpu_model
# pytest -x -v -s tests/kernels/attention/test_mla_decode_cpu.py -m cpu_model
# Note: disable Bart until supports V1
pytest -x -v -s tests/models/language/generation -m cpu_model \
--ignore=tests/models/language/generation/test_bart.py
VLLM_CPU_SGL_KERNEL=1 pytest -x -v -s tests/models/language/generation -m cpu_model \
--ignore=tests/models/language/generation/test_bart.py
pytest -x -v -s tests/models/language/generation -m cpu_model
VLLM_CPU_SGL_KERNEL=1 pytest -x -v -s tests/models/language/generation -m cpu_model
pytest -x -v -s tests/models/language/pooling -m cpu_model
pytest -x -v -s tests/models/multimodal/generation \

View File

@ -0,0 +1,191 @@
#!/bin/bash
# This script build the Ascend NPU docker image and run the offline inference inside the container.
# It serves a sanity check for compilation and basic model usage.
set -ex
# Base ubuntu image with basic ascend development libraries and python installed
VLLM_ASCEND_REPO="https://github.com/vllm-project/vllm-ascend.git"
CONFIG_FILE_REMOTE_PATH="tests/e2e/vllm_interface/vllm_test.cfg"
TEST_RUN_CONFIG_FILE="vllm_test.cfg"
VLLM_ASCEND_TMP_DIR=
# Get the test run configuration file from the vllm-ascend repository
fetch_vllm_test_cfg() {
VLLM_ASCEND_TMP_DIR=$(mktemp -d)
# Ensure that the temporary directory is cleaned up when an exception occurs during configuration file retrieval
cleanup() {
rm -rf "${VLLM_ASCEND_TMP_DIR}"
}
trap cleanup EXIT
GIT_TRACE=1 git clone -v --depth 1 "${VLLM_ASCEND_REPO}" "${VLLM_ASCEND_TMP_DIR}"
if [ ! -f "${VLLM_ASCEND_TMP_DIR}/${CONFIG_FILE_REMOTE_PATH}" ]; then
echo "Error: file '${CONFIG_FILE_REMOTE_PATH}' does not exist in the warehouse" >&2
exit 1
fi
# If the file already exists locally, just overwrite it
cp "${VLLM_ASCEND_TMP_DIR}/${CONFIG_FILE_REMOTE_PATH}" "${TEST_RUN_CONFIG_FILE}"
echo "Copied ${CONFIG_FILE_REMOTE_PATH} to ${TEST_RUN_CONFIG_FILE}"
# Since the trap will be overwritten later, and when it is executed here, the task of cleaning up resources
# when the trap is abnormal has been completed, so the temporary resources are manually deleted here.
rm -rf "${VLLM_ASCEND_TMP_DIR}"
trap - EXIT
}
# Downloads test run configuration file from a remote URL.
# Loads the configuration into the current script environment.
get_config() {
if [ ! -f "${TEST_RUN_CONFIG_FILE}" ]; then
echo "Error: file '${TEST_RUN_CONFIG_FILE}' does not exist in the warehouse" >&2
exit 1
fi
source "${TEST_RUN_CONFIG_FILE}"
echo "Base docker image name that get from configuration: ${BASE_IMAGE_NAME}"
return 0
}
# get test running configuration.
fetch_vllm_test_cfg
get_config
# Check if the function call was successful. If not, exit the script.
if [ $? -ne 0 ]; then
exit 1
fi
image_name="npu/vllm-ci:${BUILDKITE_COMMIT}_${EPOCHSECONDS}"
container_name="npu_${BUILDKITE_COMMIT}_$(tr -dc A-Za-z0-9 < /dev/urandom | head -c 10; echo)"
# BUILDKITE_AGENT_NAME format is {hostname}-{agent_idx}-{npu_card_num}cards
agent_idx=$(echo "${BUILDKITE_AGENT_NAME}" | awk -F'-' '{print $(NF-1)}')
echo "agent_idx: ${agent_idx}"
builder_name="cachebuilder${agent_idx}"
builder_cache_dir="/mnt/docker-cache${agent_idx}"
mkdir -p ${builder_cache_dir}
# Try building the docker image
cat <<EOF | DOCKER_BUILDKIT=1 docker build \
--add-host cache-service-vllm.nginx-pypi-cache.svc.cluster.local:${PYPI_CACHE_HOST} \
--builder ${builder_name} --cache-from type=local,src=${builder_cache_dir} \
--cache-to type=local,dest=${builder_cache_dir},mode=max \
--progress=plain --load -t ${image_name} -f - .
FROM ${BASE_IMAGE_NAME}
# Define environments
ENV DEBIAN_FRONTEND=noninteractive
RUN pip config set global.index-url http://cache-service-vllm.nginx-pypi-cache.svc.cluster.local:${PYPI_CACHE_PORT}/pypi/simple && \
pip config set global.trusted-host cache-service-vllm.nginx-pypi-cache.svc.cluster.local && \
apt-get update -y && \
apt-get install -y python3-pip git vim wget net-tools gcc g++ cmake libnuma-dev && \
rm -rf /var/cache/apt/* && \
rm -rf /var/lib/apt/lists/*
# Install for pytest to make the docker build cache layer always valid
RUN --mount=type=cache,target=/root/.cache/pip \
pip install pytest>=6.0 modelscope
WORKDIR /workspace/vllm
# Install vLLM dependencies in advance. Effect: As long as common.txt remains unchanged, the docker cache layer will be valid.
COPY requirements/common.txt /workspace/vllm/requirements/common.txt
RUN --mount=type=cache,target=/root/.cache/pip \
pip install -r requirements/common.txt
COPY . .
# Install vLLM
RUN --mount=type=cache,target=/root/.cache/pip \
VLLM_TARGET_DEVICE="empty" python3 -m pip install -v -e /workspace/vllm/ --extra-index https://download.pytorch.org/whl/cpu/ && \
python3 -m pip uninstall -y triton
# Install vllm-ascend
WORKDIR /workspace
ARG VLLM_ASCEND_REPO=https://github.com/vllm-project/vllm-ascend.git
ARG VLLM_ASCEND_TAG=main
RUN git config --global url."https://gh-proxy.test.osinfra.cn/https://github.com/".insteadOf "https://github.com/" && \
git clone --depth 1 \$VLLM_ASCEND_REPO --branch \$VLLM_ASCEND_TAG /workspace/vllm-ascend
# Install vllm dependencies in advance. Effect: As long as common.txt remains unchanged, the docker cache layer will be valid.
RUN --mount=type=cache,target=/root/.cache/pip \
pip install -r /workspace/vllm-ascend/requirements.txt
RUN --mount=type=cache,target=/root/.cache/pip \
export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi && \
source /usr/local/Ascend/ascend-toolkit/set_env.sh && \
source /usr/local/Ascend/nnal/atb/set_env.sh && \
export LD_LIBRARY_PATH=\$LD_LIBRARY_PATH:/usr/local/Ascend/ascend-toolkit/latest/`uname -i`-linux/devlib && \
python3 -m pip install -v -e /workspace/vllm-ascend/ --extra-index https://download.pytorch.org/whl/cpu/
ENV VLLM_WORKER_MULTIPROC_METHOD=spawn
ENV VLLM_USE_MODELSCOPE=True
WORKDIR /workspace/vllm-ascend
CMD ["/bin/bash"]
EOF
# Setup cleanup
remove_docker_container() {
docker rm -f "${container_name}" || true;
docker image rm -f "${image_name}" || true;
docker system prune -f || true;
}
trap remove_docker_container EXIT
# Generate corresponding --device args based on BUILDKITE_AGENT_NAME
# Ascend NPU BUILDKITE_AGENT_NAME format is {hostname}-{agent_idx}-{npu_card_num}cards, and agent_idx starts from 1.
# e.g. atlas-a2-001-1-2cards means this is the 1-th agent on atlas-a2-001 host, and it has 2 NPU cards.
# returns --device /dev/davinci0 --device /dev/davinci1
parse_and_gen_devices() {
local input="$1"
local index cards_num
if [[ "$input" =~ ([0-9]+)-([0-9]+)cards$ ]]; then
index="${BASH_REMATCH[1]}"
cards_num="${BASH_REMATCH[2]}"
else
echo "parse error" >&2
return 1
fi
local devices=""
local i=0
while (( i < cards_num )); do
local dev_idx=$(((index - 1)*cards_num + i ))
devices="$devices --device /dev/davinci${dev_idx}"
((i++))
done
# trim leading space
devices="${devices#"${devices%%[![:space:]]*}"}"
# Output devices: assigned to the caller variable
printf '%s' "$devices"
}
devices=$(parse_and_gen_devices "${BUILDKITE_AGENT_NAME}") || exit 1
# Run the image and execute the Out-Of-Tree (OOT) platform interface test case on Ascend NPU hardware.
# This test checks whether the OOT platform interface is functioning properly in conjunction with
# the hardware plugin vllm-ascend.
model_cache_dir=/mnt/modelscope${agent_idx}
mkdir -p ${model_cache_dir}
docker run \
${devices} \
--device /dev/davinci_manager \
--device /dev/devmm_svm \
--device /dev/hisi_hdc \
-v /usr/local/dcmi:/usr/local/dcmi \
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
-v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
-v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
-v /etc/ascend_install.info:/etc/ascend_install.info \
-v ${model_cache_dir}:/root/.cache/modelscope \
--entrypoint="" \
--name "${container_name}" \
"${image_name}" \
bash -c '
set -e
pytest -v -s tests/e2e/vllm_interface/
'

View File

@ -62,7 +62,7 @@ echo "--- Installing Python dependencies ---"
python3 -m pip install --progress-bar off git+https://github.com/thuml/depyf.git \
&& python3 -m pip install --progress-bar off pytest pytest-asyncio tpu-info \
&& python3 -m pip install --progress-bar off "lm-eval @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d" \
&& python3 -m pip install --progress-bar off hf-transfer
&& python3 -m pip install --progress-bar off hf-transfer tblib==3.1.0
echo "--- Python dependencies installed ---"
export VLLM_USE_V1=1
export VLLM_XLA_CHECK_RECOMPILATION=1

View File

@ -62,7 +62,7 @@ echo "--- Installing Python dependencies ---"
python3 -m pip install --progress-bar off git+https://github.com/thuml/depyf.git \
&& python3 -m pip install --progress-bar off pytest pytest-asyncio tpu-info \
&& python3 -m pip install --progress-bar off "lm-eval @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d" \
&& python3 -m pip install --progress-bar off hf-transfer
&& python3 -m pip install --progress-bar off hf-transfer tblib==3.1.0
echo "--- Python dependencies installed ---"
export VLLM_USE_V1=1
export VLLM_XLA_CHECK_RECOMPILATION=1

View File

@ -35,16 +35,15 @@ docker run \
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 -O3 -O.cudagraph_mode=NONE
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend ray
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend mp
VLLM_ATTENTION_BACKEND=TRITON_ATTN_VLLM_V1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager
VLLM_ATTENTION_BACKEND=TRITON_ATTN python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager
cd tests
pytest -v -s v1/core
pytest -v -s v1/engine
pytest -v -s v1/sample --ignore=v1/sample/test_logprobs.py --ignore=v1/sample/test_logprobs_e2e.py
pytest -v -s v1/worker --ignore=v1/worker/test_gpu_model_runner.py
pytest -v -s v1/structured_output
pytest -v -s v1/spec_decode --ignore=v1/spec_decode/test_max_len.py --ignore=v1/spec_decode/test_eagle.py --ignore=v1/spec_decode/test_tree_attention.py
pytest -v -s v1/spec_decode --ignore=v1/spec_decode/test_max_len.py --ignore=v1/spec_decode/test_tree_attention.py
pytest -v -s v1/kv_connector/unit --ignore=v1/kv_connector/unit/test_multi_connector.py --ignore=v1/kv_connector/unit/test_nixl_connector.py --ignore=v1/kv_connector/unit/test_shared_storage_connector.py
pytest -v -s v1/test_metrics
pytest -v -s v1/test_serial_utils.py
pytest -v -s v1/test_utils.py
pytest -v -s v1/test_metrics_reader.py
'

View File

@ -18,7 +18,7 @@ vllm bench throughput --input-len 256 --output-len 256 --output-json throughput_
bench_throughput_exit_code=$?
# run server-based benchmarks and upload the result to buildkite
python3 -m vllm.entrypoints.openai.api_server --model meta-llama/Llama-2-7b-chat-hf &
vllm serve meta-llama/Llama-2-7b-chat-hf &
server_pid=$!
wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/resolve/main/ShareGPT_V3_unfiltered_cleaned_split.json

View File

@ -0,0 +1,59 @@
#!/bin/bash
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# Setup script for Prime-RL integration tests
# This script prepares the environment for running Prime-RL tests with nightly vLLM
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
REPO_ROOT="$(cd "${SCRIPT_DIR}/../.." && pwd)"
PRIME_RL_REPO="https://github.com/PrimeIntellect-ai/prime-rl.git"
PRIME_RL_DIR="${REPO_ROOT}/prime-rl"
echo "Setting up Prime-RL integration test environment..."
# Clean up any existing Prime-RL directory
if [ -d "${PRIME_RL_DIR}" ]; then
echo "Removing existing Prime-RL directory..."
rm -rf "${PRIME_RL_DIR}"
fi
# Install UV if not available
if ! command -v uv &> /dev/null; then
echo "Installing UV package manager..."
curl -LsSf https://astral.sh/uv/install.sh | sh
source $HOME/.local/bin/env
fi
# Clone Prime-RL repository at specific branch for reproducible tests
PRIME_RL_BRANCH="integ-vllm-main"
echo "Cloning Prime-RL repository at branch: ${PRIME_RL_BRANCH}..."
git clone --branch "${PRIME_RL_BRANCH}" --single-branch "${PRIME_RL_REPO}" "${PRIME_RL_DIR}"
cd "${PRIME_RL_DIR}"
echo "Setting up UV project environment..."
export UV_PROJECT_ENVIRONMENT=/usr/local
ln -s /usr/bin/python3 /usr/local/bin/python
# Remove vllm pin from pyproject.toml
echo "Removing vllm pin from pyproject.toml..."
sed -i '/vllm==/d' pyproject.toml
# Sync Prime-RL dependencies
echo "Installing Prime-RL dependencies..."
uv sync --inexact && uv sync --inexact --all-extras
# Verify installation
echo "Verifying installations..."
uv run python -c "import vllm; print(f'vLLM version: {vllm.__version__}')"
uv run python -c "import prime_rl; print('Prime-RL imported successfully')"
echo "Prime-RL integration test environment setup complete!"
echo "Running Prime-RL integration tests..."
export WANDB_MODE=offline # this makes this test not require a WANDB_API_KEY
uv run pytest -vs tests/integration/test_rl.py -m gpu
echo "Prime-RL integration tests completed!"

View File

@ -6,24 +6,28 @@
# to generate the final pipeline yaml file.
# Documentation
# label(str): the name of the test. emoji allowed.
# fast_check(bool): whether to run this on each commit on fastcheck pipeline.
# torch_nightly(bool): whether to run this on vllm against torch nightly pipeline.
# fast_check_only(bool): run this test on fastcheck pipeline only
# optional(bool): never run this test by default (i.e. need to unblock manually) unless it's scheduled nightly run.
# label(str): the name of the test. emojis allowed.
# fast_check(bool): whether to run this on each commit on the fastcheck pipeline.
# torch_nightly(bool): whether to run this on vllm against the torch nightly pipeline.
# fast_check_only(bool): run this test on the fastcheck pipeline only
# optional(bool): never run this test by default (i.e. need to unblock manually) unless it's a scheduled nightly run.
# soft_fail(bool): allow this step to fail without failing the entire pipeline (useful for flaky or experimental tests).
# command(str): the single command to run for tests. incompatible with commands.
# commands(list): the list of commands to run for test. incompatbile with command.
# mirror_hardwares(list): the list of hardwares to run the test on as well. currently only supports [amd]
# gpu(str): override the GPU selection for the test. default is on L4 GPUs. currently only supports a100
# num_gpus(int): override the number of GPUs for the test. default to 1 GPU. currently support 2,4.
# num_nodes(int): whether to simulate multi-node setup by launch multiple containers on one host,
# in this case, commands must be specified. the first command runs on first host, the second
# commands(list): the list of commands to run for the test. incompatible with command.
# mirror_hardwares(list): the list of hardware to run the test on as well. currently only supports [amdexperimental]
# gpu(str): override the GPU selection for the test. default is L4 GPUs. supports a100, b200, h200
# num_gpus(int): override the number of GPUs for the test. defaults to 1 GPU. currently supports 2,4.
# num_nodes(int): whether to simulate multi-node setup by launching multiple containers on one host,
# in this case, commands must be specified. the first command runs on the first host, the second
# command runs on the second host.
# working_dir(str): specify the place where command should execute, default to /vllm-workspace/tests
# source_file_dependencies(list): the list of prefix to opt-in the test for, if empty, the test will always run.
# timeout_in_minutes(int): sets a timeout for the step in minutes. if not specified, uses the default timeout.
# parallelism(int): number of parallel jobs to run for this step. enables test sharding using $$BUILDKITE_PARALLEL_JOB
# and $$BUILDKITE_PARALLEL_JOB_COUNT environment variables.
# working_dir(str): specify the place where the command should execute, default to /vllm-workspace/tests
# source_file_dependencies(list): the list of prefixes to opt-in the test for, if empty, the test will always run.
# When adding a test
# - If the test belong to an existing group, add it there
# - If the test belongs to an existing group, add it there
# - If the test is short, add to any existing step
# - If the test takes more than 10min, then it is okay to create a new step.
# Note that all steps execute in parallel.
@ -46,19 +50,28 @@ steps:
mirror_hardwares: [amdexperimental]
source_file_dependencies:
- vllm/
- tests/multimodal
- tests/utils_
commands:
- pytest -v -s -m 'not cpu_test' multimodal
- pytest -v -s utils_
- label: Async Engine, Inputs, Utils, Worker Test (CPU) # 4 mins
timeout_in_minutes: 10
source_file_dependencies:
- vllm/
- tests/test_inputs.py
- tests/test_outputs.py
- tests/multimodal
- tests/utils_
- tests/standalone_tests/lazy_imports.py
- tests/transformers_utils
no_gpu: true
commands:
- python3 standalone_tests/lazy_imports.py
- pytest -v -s test_inputs.py
- pytest -v -s test_outputs.py
- pytest -v -s multimodal
- pytest -v -s utils_ # Utils
- pytest -v -s transformers_utils # transformers_utils
- pytest -v -s -m 'cpu_test' multimodal
- pytest -v -s transformers_utils
- label: Python-only Installation Test # 10min
timeout_in_minutes: 20
@ -110,7 +123,7 @@ steps:
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
- pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py
- pytest -v -s entrypoints/llm/test_generate.py # it needs a clean process
- VLLM_USE_V1=0 pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
- pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
- label: Entrypoints Integration Test (API Server) # 100min
timeout_in_minutes: 130
@ -148,7 +161,6 @@ steps:
num_gpus: 4
source_file_dependencies:
- vllm/distributed/
- vllm/core/
- tests/distributed/test_utils
- tests/distributed/test_pynccl
- tests/distributed/test_events
@ -156,28 +168,34 @@ steps:
- examples/offline_inference/rlhf.py
- examples/offline_inference/rlhf_colocate.py
- tests/examples/offline_inference/data_parallel.py
- tests/v1/test_async_llm_dp.py
- tests/v1/test_external_lb_dp.py
- tests/v1/test_internal_lb_dp.py
- tests/v1/test_hybrid_lb_dp.py
- tests/v1/distributed
- tests/v1/engine/test_engine_core_client.py
- tests/distributed/test_symm_mem_allreduce.py
commands:
# test with tp=2 and external_dp=2
- VLLM_USE_V1=0 torchrun --nproc-per-node=4 distributed/test_torchrun_example.py
# test with torchrun tp=2 and external_dp=2
- torchrun --nproc-per-node=4 distributed/test_torchrun_example.py
# test with tp=2 and pp=2
# test with torchrun tp=2 and pp=2
- PP_SIZE=2 torchrun --nproc-per-node=4 distributed/test_torchrun_example.py
# test with torchrun tp=4 and dp=1
- TP_SIZE=4 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py
# test with torchrun tp=2, pp=2 and dp=1
- PP_SIZE=2 TP_SIZE=2 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py
# test with torchrun tp=1 and dp=4 with ep
- DP_SIZE=4 ENABLE_EP=1 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py
# test with torchrun tp=2 and dp=2 with ep
- TP_SIZE=2 DP_SIZE=2 ENABLE_EP=1 torchrun --nproc-per-node=4 distributed/test_torchrun_example_moe.py
# test with internal dp
- python3 ../examples/offline_inference/data_parallel.py --enforce-eager
- TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/test_async_llm_dp.py
- TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/test_external_lb_dp.py
- TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/test_internal_lb_dp.py
- TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/test_hybrid_lb_dp.py
- TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py
- TP_SIZE=2 DP_SIZE=2 pytest -v -s v1/distributed/test_external_lb_dp.py
- TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/distributed/test_internal_lb_dp.py
- TP_SIZE=1 DP_SIZE=4 pytest -v -s v1/distributed/test_hybrid_lb_dp.py
- pytest -v -s v1/engine/test_engine_core_client.py::test_kv_cache_events_dp
- pytest -v -s distributed/test_utils.py
- pytest -v -s compile/test_basic_correctness.py
- pytest -v -s distributed/test_pynccl.py
- pytest -v -s distributed/test_events.py
- pytest -v -s distributed/test_symm_mem_allreduce.py
# TODO: create a dedicated test section for multi-GPU example tests
# when we have multiple distributed example tests
- pushd ../examples/offline_inference
@ -278,24 +296,34 @@ steps:
- tests/v1
commands:
# split the test to avoid interference
- pytest -v -s v1/core
- pytest -v -s v1/executor
- pytest -v -s v1/kv_offload
- pytest -v -s v1/sample
- pytest -v -s v1/logits_processors
- pytest -v -s v1/worker
- pytest -v -s v1/structured_output
- pytest -v -s v1/spec_decode
- pytest -v -s v1/kv_connector/unit
- pytest -v -s v1/metrics
- pytest -v -s v1/test_serial_utils.py
- pytest -v -s v1/test_utils.py
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
- pytest -v -s -m 'not cpu_test' v1/metrics
- pytest -v -s v1/test_oracle.py
- pytest -v -s v1/test_metrics_reader.py
- pytest -v -s v1/test_request.py
# Integration test for streaming correctness (requires special branch).
- pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
- label: V1 Test others (CPU) # 5 mins
source_file_dependencies:
- vllm/
- tests/v1
no_gpu: true
commands:
# split the test to avoid interference
- pytest -v -s v1/core
- pytest -v -s v1/structured_output
- pytest -v -s v1/test_serial_utils.py
- pytest -v -s -m 'cpu_test' v1/kv_connector/unit
- pytest -v -s -m 'cpu_test' v1/metrics
- label: Examples Test # 30min
timeout_in_minutes: 45
mirror_hardwares: [amdexperimental]
@ -314,12 +342,13 @@ steps:
- python3 offline_inference/vision_language.py --seed 0
- python3 offline_inference/vision_language_pooling.py --seed 0
- python3 offline_inference/vision_language_multi_image.py --seed 0
- VLLM_USE_V1=0 python3 others/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 others/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors
- python3 others/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 others/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors
- python3 offline_inference/encoder_decoder_multimodal.py --model-type whisper --seed 0
- python3 offline_inference/basic/classify.py
- python3 offline_inference/basic/embed.py
- python3 offline_inference/basic/score.py
- VLLM_USE_V1=0 python3 offline_inference/profiling.py --model facebook/opt-125m run_num_steps --num-steps 2
- python3 offline_inference/spec_decode.py --test --method eagle --num_spec_tokens 3 --dataset-name hf --dataset-path philschmid/mt-bench --num-prompts 80 --temp 0 --top-p 1.0 --top-k -1 --tp 1 --enable-chunked-prefill --max-model-len 2048
- python3 offline_inference/spec_decode.py --test --method eagle3 --num_spec_tokens 3 --dataset-name hf --dataset-path philschmid/mt-bench --num-prompts 80 --temp 0 --top-p 1.0 --top-k -1 --tp 1 --enable-chunked-prefill --max-model-len 2048
- label: Platform Tests (CUDA) # 4min
timeout_in_minutes: 15
@ -450,29 +479,18 @@ steps:
commands:
- pytest -v -s kernels/mamba
- label: Tensorizer Test # 14min
timeout_in_minutes: 25
mirror_hardwares: [amdexperimental]
source_file_dependencies:
- vllm/model_executor/model_loader
- tests/tensorizer_loader
- tests/entrypoints/openai/test_tensorizer_entrypoint.py
commands:
- apt-get update && apt-get install -y curl libsodium23
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
- pytest -v -s tensorizer_loader
- pytest -v -s entrypoints/openai/test_tensorizer_entrypoint.py
- label: Model Executor Test # 7min
timeout_in_minutes: 20
- label: Model Executor Test # 23min
timeout_in_minutes: 35
mirror_hardwares: [amdexperimental]
source_file_dependencies:
- vllm/model_executor
- tests/model_executor
- tests/entrypoints/openai/test_tensorizer_entrypoint.py
commands:
- apt-get update && apt-get install -y curl libsodium23
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
- pytest -v -s model_executor
- pytest -v -s entrypoints/openai/test_tensorizer_entrypoint.py
- label: Benchmarks # 11min
timeout_in_minutes: 20
@ -507,7 +525,7 @@ steps:
# https://github.com/pytorch/ao/issues/2919, we'll have to skip new torchao tests for now
# we can only upgrade after this is resolved
- pip install --pre torchao==0.13.0.dev20250814 --index-url https://download.pytorch.org/whl/nightly/cu128
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization
- VLLM_TEST_FORCE_LOAD_FORMAT=auto pytest -v -s quantization/
- label: LM Eval Small Models # 53min
timeout_in_minutes: 75
@ -535,10 +553,17 @@ steps:
source_file_dependencies:
- vllm/
- tests/tool_use
- tests/mistral_tool_use
commands:
- pytest -v -s tool_use
- pytest -v -s mistral_tool_use
- pytest -v -s -m 'not cpu_test' tool_use
- label: OpenAI-Compatible Tool Use (CPU) # 5 mins
timeout_in_minutes: 10
source_file_dependencies:
- vllm/
- tests/tool_use
no_gpu: true
commands:
- pytest -v -s -m 'cpu_test' tool_use
##### models test #####
@ -578,13 +603,19 @@ steps:
- vllm/
- tests/models/test_transformers.py
- tests/models/test_registry.py
commands:
- pytest -v -s models/test_transformers.py models/test_registry.py
- label: Basic Models Test (Other CPU) # 5min
timeout_in_minutes: 10
torch_nightly: true
source_file_dependencies:
- vllm/
- tests/models/test_utils.py
- tests/models/test_vision.py
no_gpu: true
commands:
- pytest -v -s models/test_transformers.py \
models/test_registry.py \
models/test_utils.py \
models/test_vision.py
- pytest -v -s models/test_utils.py models/test_vision.py
- label: Language Models Tests (Standard)
timeout_in_minutes: 25
@ -754,11 +785,13 @@ steps:
commands:
- pip install --upgrade git+https://github.com/huggingface/transformers
- pytest -v -s tests/models/test_initialization.py
- pytest -v -s tests/models/test_transformers.py
- pytest -v -s tests/models/multimodal/processing/
- pytest -v -s tests/models/multimodal/test_mapping.py
- python3 examples/offline_inference/basic/chat.py
- python3 examples/offline_inference/audio_language.py --model-type whisper
- python3 examples/offline_inference/vision_language.py --model-type qwen2_5_vl
# Whisper needs spawn method to avoid deadlock
- VLLM_WORKER_MULTIPROC_METHOD=spawn python3 examples/offline_inference/audio_language.py --model-type whisper
- label: Blackwell Test # 38 min
timeout_in_minutes: 60
@ -812,7 +845,24 @@ steps:
- vllm/v1/attention/backends/flashinfer.py
commands:
- uv pip install --system 'gpt-oss[eval]==0.0.5'
- pytest -s -v tests/evals/gpt_oss/test_gpqa_correctness.py --model openai/gpt-oss-20b --metric 0.58 --server-args '--tensor-parallel-size 2'
- pytest -s -v tests/evals/gpt_oss/test_gpqa_correctness.py --model openai/gpt-oss-20b --metric 0.58
- label: Blackwell Quantized MoE Test
timeout_in_minutes: 60
working_dir: "/vllm-workspace/"
gpu: b200
source_file_dependencies:
- tests/quantization/test_blackwell_moe.py
- vllm/model_executor/models/deepseek_v2.py
- vllm/model_executor/models/gpt_oss.py
- vllm/model_executor/models/llama4.py
- vllm/model_executor/layers/fused_moe
- vllm/model_executor/layers/quantization/compressed_tensors
- vllm/model_executor/layers/quantization/modelopt.py
- vllm/model_executor/layers/quantization/mxfp4.py
- vllm/v1/attention/backends/flashinfer.py
commands:
- pytest -s -v tests/quantization/test_blackwell_moe.py
##### 1 GPU test #####
##### multi gpus test #####
@ -856,47 +906,58 @@ steps:
- NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed'
- python3 ../examples/offline_inference/data_parallel.py --dp-size=2 --tp-size=1 --node-size=2 --node-rank=1 --master-addr=192.168.10.10 --master-port=12345 --enforce-eager --trust-remote-code
- label: Distributed Tests (2 GPUs) # 110min
timeout_in_minutes: 150
- label: Distributed Tests (2 GPUs) # 68min
timeout_in_minutes: 90
mirror_hardwares: [amdexperimental]
working_dir: "/vllm-workspace/tests"
num_gpus: 2
source_file_dependencies:
- vllm/compilation/
- vllm/distributed/
- vllm/engine/
- vllm/executor/
- vllm/model_executor/models/
- tests/distributed/
- vllm/compilation
- vllm/worker/worker_base.py
- vllm/worker/worker.py
- vllm/worker/model_runner.py
- entrypoints/llm/test_collective_rpc.py
- tests/v1/test_async_llm_dp.py
- tests/v1/test_external_lb_dp.py
- tests/v1/entrypoints/openai/test_multi_api_servers.py
- vllm/v1/engine/
- vllm/v1/worker/
- tests/compile/test_basic_correctness.py
- tests/compile/test_wrapper.py
- tests/distributed/
- tests/entrypoints/llm/test_collective_rpc.py
- tests/v1/distributed
- tests/v1/entrypoints/openai/test_multi_api_servers.py
- tests/v1/shutdown
- tests/v1/worker/test_worker_memory_snapshot.py
commands:
- TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/test_async_llm_dp.py
- TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/test_external_lb_dp.py
- TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_async_llm_dp.py
- TP_SIZE=1 DP_SIZE=2 pytest -v -s v1/distributed/test_external_lb_dp.py
- DP_SIZE=2 pytest -v -s v1/entrypoints/openai/test_multi_api_servers.py
- pytest -v -s entrypoints/llm/test_collective_rpc.py
- pytest -v -s ./compile/test_basic_correctness.py
- pytest -v -s ./compile/test_wrapper.py
- VLLM_TEST_SAME_HOST=1 torchrun --nproc-per-node=4 distributed/test_same_node.py | grep 'Same node test passed'
- pytest -v -s distributed/test_sequence_parallel.py
- CUDA_VISIBLE_DEVICES=0,1 pytest -v -s v1/shutdown
- pytest -v -s v1/worker/test_worker_memory_snapshot.py
- label: Distributed Model Tests (2 GPUs) # 37min
timeout_in_minutes: 50
mirror_hardwares: [amdexperimental]
working_dir: "/vllm-workspace/tests"
num_gpus: 2
source_file_dependencies:
- vllm/model_executor/model_loader/sharded_state_loader.py
- vllm/model_executor/models/
- tests/basic_correctness/
- tests/model_executor/model_loader/test_sharded_state_loader.py
- tests/models/
commands:
- TARGET_TEST_SUITE=L4 pytest basic_correctness/ -v -s -m 'distributed(num_gpus=2)'
- CUDA_VISIBLE_DEVICES=0,1 pytest -v -s model_executor/model_loader/test_sharded_state_loader.py
# Avoid importing model tests that cause CUDA reinitialization error
- pytest models/test_transformers.py -v -s -m 'distributed(num_gpus=2)'
- pytest models/language -v -s -m 'distributed(num_gpus=2)'
- pytest models/multimodal -v -s -m 'distributed(num_gpus=2)' --ignore models/multimodal/generation/test_whisper.py
- VLLM_WORKER_MULTIPROC_METHOD=spawn pytest models/multimodal/generation/test_whisper.py -v -s -m 'distributed(num_gpus=2)'
# test sequence parallel
- pytest -v -s distributed/test_sequence_parallel.py
# this test fails consistently.
# TODO: investigate and fix
- VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s test_sharded_state_loader.py
- CUDA_VISIBLE_DEVICES=0,1 pytest -v -s v1/shutdown
- pytest -v -s models/multimodal/generation/test_maverick.py
- label: Plugin Tests (2 GPUs) # 40min
timeout_in_minutes: 60
@ -1030,3 +1091,16 @@ steps:
num_gpus: 2
commands:
- pytest -v -s tests/distributed/test_context_parallel.py
- pytest -v -s tests/distributed/test_nccl_symm_mem_allreduce.py
##### RL Integration Tests #####
- label: Prime-RL Integration Test # 15min
timeout_in_minutes: 30
optional: true
num_gpus: 2
working_dir: "/vllm-workspace"
source_file_dependencies:
- vllm/
- .buildkite/scripts/run-prime-rl-test.sh
commands:
- bash .buildkite/scripts/run-prime-rl-test.sh

27
.github/CODEOWNERS vendored
View File

@ -4,19 +4,14 @@
# This lists cover the "core" components of vLLM that require careful review
/vllm/attention @LucasWilkinson
/vllm/attention/backends/abstract.py @WoosukKwon @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
/vllm/core @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
/vllm/engine/llm_engine.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
/vllm/executor/executor_base.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill @22quinn
/vllm/worker/worker_base.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill @22quinn
/vllm/worker/worker.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill
/vllm/model_executor/layers/fused_moe @mgoin
/vllm/model_executor/layers/sampler.py @zhuohan123 @youkaichao @alexm-redhat @comaniac @njhill @NickLucche
/vllm/model_executor/layers/quantization @mgoin @robertgshaw2-redhat @tlrmchlsmth @yewentao256
/vllm/model_executor/layers/mamba @tdoublep
/vllm/model_executor/model_loader @22quinn
/vllm/multimodal @DarkLight1337 @ywang96 @NickLucche
/vllm/v1/attention @LucasWilkinson
/vllm/v1/sample @22quinn @houseroad
/vllm/vllm_flash_attn @LucasWilkinson
/vllm/lora @jeejeelee
/vllm/reasoning @aarnphm @chaunceyjiang
@ -31,11 +26,13 @@ CMakeLists.txt @tlrmchlsmth @LucasWilkinson
# vLLM V1
/vllm/v1 @WoosukKwon @robertgshaw2-redhat @njhill @ywang96 @comaniac @alexm-redhat
/vllm/v1/structured_output @mgoin @russellb @aarnphm @benchislett
/vllm/v1/spec_decode @benchislett @luccafong
/vllm/v1/attention @LucasWilkinson
/vllm/v1/attention/backends/flashinfer.py @mgoin
/vllm/v1/attention/backends/triton_attn.py @tdoublep
/vllm/v1/core @WoosukKwon @robertgshaw2-redhat @njhill @ywang96 @comaniac @alexm-redhat @heheda12345 @ApostaC
/vllm/v1/sample @22quinn @houseroad @njhill
/vllm/v1/spec_decode @benchislett @luccafong
/vllm/v1/structured_output @mgoin @russellb @aarnphm @benchislett
/vllm/v1/kv_cache_interface.py @heheda12345
/vllm/v1/offloading @ApostaC
@ -57,7 +54,7 @@ CMakeLists.txt @tlrmchlsmth @LucasWilkinson
/tests/weight_loading @mgoin @youkaichao @yewentao256
/tests/lora @jeejeelee
/tests/models/language/generation/test_hybrid.py @tdoublep
/tests/v1/kv_connector/nixl_integration @NickLucche
/tests/v1/kv_connector/nixl_integration @NickLucche
/tests/v1/kv_connector @ApostaC
/tests/v1/offloading @ApostaC
@ -66,18 +63,26 @@ CMakeLists.txt @tlrmchlsmth @LucasWilkinson
/tests/models/test_transformers.py @hmellor
# Docs
/docs @hmellor
/docs/mkdocs @hmellor
/docs/**/*.yml @hmellor
/requirements/docs.txt @hmellor
.readthedocs.yaml @hmellor
mkdocs.yaml @hmellor
# Linting
.markdownlint.yaml @hmellor
.pre-commit-config.yaml @hmellor
/tools/pre_commit @hmellor
# CPU
/vllm/v1/worker/^cpu @bigPYJ1151
/vllm/v1/worker/cpu* @bigPYJ1151
/csrc/cpu @bigPYJ1151
/vllm/platforms/cpu.py @bigPYJ1151
/cmake/cpu_extension.cmake @bigPYJ1151
/docker/Dockerfile.cpu @bigPYJ1151
# Intel GPU
/vllm/v1/worker/^xpu @jikunshang
/vllm/v1/worker/xpu* @jikunshang
/vllm/platforms/xpu.py @jikunshang
/docker/Dockerfile.xpu @jikunshang

View File

@ -43,10 +43,6 @@ body:
Any other things you would like to mention.
validations:
required: false
- type: markdown
attributes:
value: >
Thanks for contributing 🎉! The vLLM core team hosts a biweekly RFC review session at 9:30AM Pacific Time, while most RFCs can be discussed online, you can optionally sign up for a slot to discuss your RFC online [here](https://docs.google.com/document/d/1CiLVBZeIVfR7_PNAKVSusxpceywkoOOB78qoWqHvSZc/edit).
- type: checkboxes
id: askllm
attributes:

33
.github/mergify.yml vendored
View File

@ -2,6 +2,7 @@ pull_request_rules:
- name: label-documentation
description: Automatically apply documentation label
conditions:
- label != stale
- or:
- files~=^[^/]+\.md$
- files~=^docs/
@ -14,6 +15,7 @@ pull_request_rules:
- name: label-ci-build
description: Automatically apply ci/build label
conditions:
- label != stale
- or:
- files~=^\.github/
- files~=\.buildkite/
@ -30,6 +32,7 @@ pull_request_rules:
- name: label-deepseek
description: Automatically apply deepseek label
conditions:
- label != stale
- or:
- files~=^examples/.*deepseek.*\.py
- files~=^tests/.*deepseek.*\.py
@ -46,6 +49,7 @@ pull_request_rules:
- name: label-frontend
description: Automatically apply frontend label
conditions:
- label != stale
- files~=^vllm/entrypoints/
actions:
label:
@ -55,6 +59,7 @@ pull_request_rules:
- name: label-llama
description: Automatically apply llama label
conditions:
- label != stale
- or:
- files~=^examples/.*llama.*\.py
- files~=^tests/.*llama.*\.py
@ -70,6 +75,7 @@ pull_request_rules:
- name: label-multi-modality
description: Automatically apply multi-modality label
conditions:
- label != stale
- or:
- files~=^vllm/multimodal/
- files~=^tests/multimodal/
@ -83,6 +89,7 @@ pull_request_rules:
- name: label-new-model
description: Automatically apply new-model label
conditions:
- label != stale
- and:
- files~=^vllm/model_executor/models/
- files=vllm/model_executor/models/registry.py
@ -94,6 +101,7 @@ pull_request_rules:
- name: label-performance
description: Automatically apply performance label
conditions:
- label != stale
- or:
- files~=^benchmarks/
- files~=^vllm/benchmarks/
@ -107,6 +115,7 @@ pull_request_rules:
- name: label-qwen
description: Automatically apply qwen label
conditions:
- label != stale
- or:
- files~=^examples/.*qwen.*\.py
- files~=^tests/.*qwen.*\.py
@ -121,6 +130,7 @@ pull_request_rules:
- name: label-gpt-oss
description: Automatically apply gpt-oss label
conditions:
- label != stale
- or:
- files~=^examples/.*gpt[-_]?oss.*\.py
- files~=^tests/.*gpt[-_]?oss.*\.py
@ -142,6 +152,7 @@ pull_request_rules:
- name: label-rocm
description: Automatically apply rocm label
conditions:
- label != stale
- or:
- files~=^csrc/rocm/
- files~=^docker/Dockerfile.rocm
@ -162,6 +173,7 @@ pull_request_rules:
- name: label-structured-output
description: Automatically apply structured-output label
conditions:
- label != stale
- or:
- files~=^benchmarks/structured_schemas/
- files=benchmarks/benchmark_serving_structured_output.py
@ -181,6 +193,7 @@ pull_request_rules:
- name: label-speculative-decoding
description: Automatically apply speculative-decoding label
conditions:
- label != stale
- or:
- files~=^vllm/v1/spec_decode/
- files~=^tests/v1/spec_decode/
@ -196,6 +209,7 @@ pull_request_rules:
- name: label-v1
description: Automatically apply v1 label
conditions:
- label != stale
- or:
- files~=^vllm/v1/
- files~=^tests/v1/
@ -208,6 +222,7 @@ pull_request_rules:
description: Automatically apply tpu label
# Keep this list in sync with `label-tpu-remove` conditions
conditions:
- label != stale
- or:
- files~=tpu.py
- files~=_tpu
@ -223,6 +238,7 @@ pull_request_rules:
description: Automatically remove tpu label
# Keep this list in sync with `label-tpu` conditions
conditions:
- label != stale
- and:
- -files~=tpu.py
- -files~=_tpu
@ -237,9 +253,9 @@ pull_request_rules:
- name: label-tool-calling
description: Automatically add tool-calling label
conditions:
- label != stale
- or:
- files~=^tests/tool_use/
- files~=^tests/mistral_tool_use/
- files~=^tests/entrypoints/openai/tool_parsers/
- files=tests/entrypoints/openai/test_chat_with_tool_reasoning.py
- files~=^vllm/entrypoints/openai/tool_parsers/
@ -256,8 +272,9 @@ pull_request_rules:
- name: ping author on conflicts and add 'needs-rebase' label
conditions:
- conflict
- -closed
- label != stale
- conflict
- -closed
actions:
label:
add:
@ -271,10 +288,12 @@ pull_request_rules:
- name: assign reviewer for tensorizer changes
conditions:
- label != stale
- or:
- files~=^vllm/model_executor/model_loader/tensorizer.py
- files~=^vllm/model_executor/model_loader/tensorizer_loader.py
- files~=^tests/entrypoints/openai/test_tensorizer_entrypoint.py
- files~=^tests/tensorizer_loader/
- files~=^tests/model_executor/model_loader/tensorizer_loader/
actions:
assign:
users:
@ -282,6 +301,7 @@ pull_request_rules:
- name: assign reviewer for modelopt changes
conditions:
- label != stale
- or:
- files~=^vllm/model_executor/layers/quantization/modelopt\.py$
- files~=^vllm/model_executor/layers/quantization/__init__\.py$
@ -296,8 +316,8 @@ pull_request_rules:
- name: remove 'needs-rebase' label when conflict is resolved
conditions:
- -conflict
- -closed
- -conflict
- -closed
actions:
label:
remove:
@ -306,6 +326,7 @@ pull_request_rules:
- name: label-kv-connector
description: Automatically apply kv-connector label
conditions:
- label != stale
- or:
- files~=^examples/online_serving/disaggregated[^/]*/.*
- files~=^examples/offline_inference/disaggregated[^/]*/.*

View File

@ -49,7 +49,7 @@ repos:
rev: 0.6.17
hooks:
- id: pip-compile
args: [requirements/test.in, -o, requirements/test.txt, --index-strategy, unsafe-best-match, --torch-backend, cu128]
args: [requirements/test.in, -o, requirements/test.txt, --index-strategy, unsafe-best-match, --torch-backend, cu128, --python-platform, x86_64-manylinux_2_28]
files: ^requirements/test\.(in|txt)$
- repo: local
hooks:
@ -60,38 +60,32 @@ repos:
files: ^requirements/test\.(in|txt)$
- id: mypy-local
name: Run mypy for local Python installation
entry: tools/mypy.sh 0 "local"
language: python
types: [python]
additional_dependencies: &mypy_deps [mypy==1.11.1, types-cachetools, types-setuptools, types-PyYAML, types-requests, pydantic]
entry: python tools/pre_commit/mypy.py 0 "local"
stages: [pre-commit] # Don't run in CI
<<: &mypy_common
language: python
types_or: [python, pyi]
require_serial: true
additional_dependencies: [mypy==1.11.1, regex, types-cachetools, types-setuptools, types-PyYAML, types-requests, types-torch, pydantic]
- id: mypy-3.9 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
name: Run mypy for Python 3.9
entry: tools/mypy.sh 1 "3.9"
language: python
types: [python]
additional_dependencies: *mypy_deps
entry: python tools/pre_commit/mypy.py 1 "3.9"
<<: *mypy_common
stages: [manual] # Only run in CI
- id: mypy-3.10 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
name: Run mypy for Python 3.10
entry: tools/mypy.sh 1 "3.10"
language: python
types: [python]
additional_dependencies: *mypy_deps
entry: python tools/pre_commit/mypy.py 1 "3.10"
<<: *mypy_common
stages: [manual] # Only run in CI
- id: mypy-3.11 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
name: Run mypy for Python 3.11
entry: tools/mypy.sh 1 "3.11"
language: python
types: [python]
additional_dependencies: *mypy_deps
entry: python tools/pre_commit/mypy.py 1 "3.11"
<<: *mypy_common
stages: [manual] # Only run in CI
- id: mypy-3.12 # TODO: Use https://github.com/pre-commit/mirrors-mypy when mypy setup is less awkward
name: Run mypy for Python 3.12
entry: tools/mypy.sh 1 "3.12"
language: python
types: [python]
additional_dependencies: *mypy_deps
entry: python tools/pre_commit/mypy.py 1 "3.12"
<<: *mypy_common
stages: [manual] # Only run in CI
- id: shellcheck
name: Lint shell scripts
@ -155,11 +149,10 @@ repos:
additional_dependencies: [regex]
- id: check-pickle-imports
name: Prevent new pickle/cloudpickle imports
entry: python tools/check_pickle_imports.py
entry: python tools/pre_commit/check_pickle_imports.py
language: python
types: [python]
pass_filenames: false
additional_dependencies: [pathspec, regex]
additional_dependencies: [regex]
- id: validate-config
name: Validate configuration has default values and that each field has a docstring
entry: python tools/validate_config.py

View File

@ -13,6 +13,7 @@ build:
mkdocs:
configuration: mkdocs.yaml
fail_on_warning: true
# Optionally declare the Python requirements required to build your docs
python:

View File

@ -37,7 +37,7 @@ install(CODE "set(CMAKE_INSTALL_LOCAL_ONLY TRUE)" ALL_COMPONENTS)
set(PYTHON_SUPPORTED_VERSIONS "3.9" "3.10" "3.11" "3.12" "3.13")
# Supported AMD GPU architectures.
set(HIP_SUPPORTED_ARCHS "gfx906;gfx908;gfx90a;gfx942;gfx950;gfx1030;gfx1100;gfx1101;gfx1200;gfx1201")
set(HIP_SUPPORTED_ARCHS "gfx906;gfx908;gfx90a;gfx942;gfx950;gfx1030;gfx1100;gfx1101;gfx1200;gfx1201;gfx1150;gfx1151")
#
# Supported/expected torch versions for CUDA/ROCm.
@ -86,6 +86,9 @@ find_package(Torch REQUIRED)
# Supported NVIDIA architectures.
# This check must happen after find_package(Torch) because that's when CMAKE_CUDA_COMPILER_VERSION gets defined
if(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 13.0)
set(CUDA_SUPPORTED_ARCHS "7.5;8.0;8.6;8.7;8.9;9.0;10.0;11.0;12.0")
elseif(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.8)
set(CUDA_SUPPORTED_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;12.0")
else()
@ -175,6 +178,15 @@ if(NVCC_THREADS AND VLLM_GPU_LANG STREQUAL "CUDA")
list(APPEND VLLM_GPU_FLAGS "--threads=${NVCC_THREADS}")
endif()
#
# Set compression mode for CUDA >=13.x.
#
if(VLLM_GPU_LANG STREQUAL "CUDA" AND
DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 13.0)
list(APPEND VLLM_GPU_FLAGS "--compress-mode=size")
endif()
#
# Set CUDA include flags for CXX compiler.
#
@ -270,7 +282,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
SET(CUTLASS_ENABLE_HEADERS_ONLY ON CACHE BOOL "Enable only the header library")
# Set CUTLASS_REVISION. Used for FetchContent. Also fixes some bogus messages when building.
set(CUTLASS_REVISION "v4.0.0" CACHE STRING "CUTLASS revision to use")
set(CUTLASS_REVISION "v4.2.1" CACHE STRING "CUTLASS revision to use")
# Use the specified CUTLASS source directory for compilation if VLLM_CUTLASS_SRC_DIR is provided
if (DEFINED ENV{VLLM_CUTLASS_SRC_DIR})
@ -305,7 +317,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
"csrc/quantization/cutlass_w8a8/scaled_mm_entry.cu"
"csrc/quantization/fp4/nvfp4_quant_entry.cu"
"csrc/quantization/fp4/nvfp4_scaled_mm_entry.cu"
"csrc/quantization/fp4/nvfp4_blockwise_moe_kernel.cu"
"csrc/sparse/cutlass/sparse_scaled_mm_entry.cu"
"csrc/cutlass_extensions/common.cpp"
"csrc/quantization/fp8/per_token_group_quant.cu")
@ -440,7 +451,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
# The cutlass_scaled_mm kernels for Geforce Blackwell SM120 (c3x, i.e. CUTLASS 3.x) require
# CUDA 12.8 or later
cuda_archs_loose_intersection(SCALED_MM_ARCHS "12.0;12.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(SCALED_MM_ARCHS "12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(SCALED_MM_ARCHS "12.0a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
set(SRCS
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm120.cu"
@ -470,7 +485,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
# The cutlass_scaled_mm kernels for Blackwell SM100 (c3x, i.e. CUTLASS 3.x)
# require CUDA 12.8 or later
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
set(SRCS
"csrc/quantization/cutlass_w8a8/scaled_mm_c3x_sm100.cu"
@ -550,7 +569,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
# The nvfp4_scaled_mm_sm120 kernels for Geforce Blackwell SM120 require
# CUDA 12.8 or later
cuda_archs_loose_intersection(FP4_ARCHS "12.0;12.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(FP4_ARCHS "12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(FP4_ARCHS "12.0a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND FP4_ARCHS)
set(SRCS
"csrc/quantization/fp4/nvfp4_quant_kernels.cu"
@ -569,7 +592,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif()
# FP4 Archs and flags
cuda_archs_loose_intersection(FP4_ARCHS "10.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(FP4_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(FP4_ARCHS "10.0a;10.1a;12.0a;12.1a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND FP4_ARCHS)
set(SRCS
"csrc/quantization/fp4/nvfp4_quant_kernels.cu"
@ -591,7 +618,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif()
# CUTLASS MLA Archs and flags
cuda_archs_loose_intersection(MLA_ARCHS "10.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(MLA_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(MLA_ARCHS "10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND MLA_ARCHS)
set(SRCS
"csrc/attention/mla/sm100_cutlass_mla_kernel.cu")
@ -635,7 +666,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif()
endif()
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0f;11.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
set(SRCS "csrc/quantization/cutlass_w8a8/moe/grouped_mm_c3x_sm100.cu")
set_gencode_flags_for_srcs(
@ -656,7 +691,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif()
# moe_data.cu is used by all CUTLASS MoE kernels.
cuda_archs_loose_intersection(CUTLASS_MOE_DATA_ARCHS "9.0a;10.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(CUTLASS_MOE_DATA_ARCHS "9.0a;10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(CUTLASS_MOE_DATA_ARCHS "9.0a;10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.3 AND CUTLASS_MOE_DATA_ARCHS)
set(SRCS "csrc/quantization/cutlass_w8a8/moe/moe_data.cu")
set_gencode_flags_for_srcs(
@ -675,7 +714,11 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif()
endif()
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 13.0)
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0f;11.0f;12.0f" "${CUDA_ARCHS}")
else()
cuda_archs_loose_intersection(SCALED_MM_ARCHS "10.0a;10.1a;10.3a;12.0a;12.1a" "${CUDA_ARCHS}")
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER_EQUAL 12.8 AND SCALED_MM_ARCHS)
set(SRCS "csrc/quantization/cutlass_w8a8/moe/blockwise_scaled_group_mm_sm100.cu")
set_gencode_flags_for_srcs(

View File

@ -21,6 +21,7 @@ Join us at the [PyTorch Conference, October 22-23](https://events.linuxfoundatio
*Latest News* 🔥
- [2025/09] We hosted [vLLM Toronto Meetup](https://luma.com/e80e0ymm) focused on tackling inference at scale and speculative decoding with speakers from NVIDIA and Red Hat! Please find the meetup slides [here](https://docs.google.com/presentation/d/1IYJYmJcu9fLpID5N5RbW_vO0XLo0CGOR14IXOjB61V8/edit?usp=sharing).
- [2025/08] We hosted [vLLM Shenzhen Meetup](https://mp.weixin.qq.com/s/k8ZBO1u2_2odgiKWH_GVTQ) focusing on the ecosystem around vLLM! Please find the meetup slides [here](https://drive.google.com/drive/folders/1Ua2SVKVSu-wp5vou_6ElraDt2bnKhiEA).
- [2025/08] We hosted [vLLM Singapore Meetup](https://www.sginnovate.com/event/vllm-sg-meet). We shared V1 updates, disaggregated serving and MLLM speedups with speakers from Embedded LLM, AMD, WekaIO, and A*STAR. Please find the meetup slides [here](https://drive.google.com/drive/folders/1ncf3GyqLdqFaB6IeB834E5TZJPLAOiXZ?usp=sharing).
- [2025/08] We hosted [vLLM Shanghai Meetup](https://mp.weixin.qq.com/s/pDmAXHcN7Iqc8sUKgJgGtg) focusing on building, developing, and integrating with vLLM! Please find the meetup slides [here](https://drive.google.com/drive/folders/1OvLx39wnCGy_WKq8SiVKf7YcxxYI3WCH).

View File

@ -103,10 +103,15 @@ start_server() {
VLLM_USE_V1=1 VLLM_SERVER_DEV_MODE=1 \
vllm serve "${common_args_array[@]}" > "$vllm_log" 2>&1 &
fi
local server_pid=$!
# wait for 10 minutes...
server_started=0
for i in {1..60}; do
# This line checks whether the server is still alive or not,
# since that we should always have permission to send signal to the server process.
kill -0 $server_pid 2> /dev/null || break
RESPONSE=$(curl -s -X GET "http://0.0.0.0:8004/health" -w "%{http_code}" -o /dev/stdout)
STATUS_CODE=$(echo "$RESPONSE" | tail -n 1)
if [[ "$STATUS_CODE" -eq 200 ]]; then
@ -118,7 +123,7 @@ start_server() {
done
if (( ! server_started )); then
echo "server did not start within 10 minutes. Please check server log at $vllm_log".
echo "server did not start within 10 minutes or crashed. Please check server log at $vllm_log".
return 1
else
return 0

View File

@ -1,17 +1,31 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import gc
import time
from unittest import mock
import numpy as np
from tabulate import tabulate
from benchmark_utils import TimeCollector
from vllm.config import ModelConfig, SpeculativeConfig, VllmConfig
from vllm.config import (
CacheConfig,
DeviceConfig,
LoadConfig,
ModelConfig,
ParallelConfig,
SchedulerConfig,
SpeculativeConfig,
VllmConfig,
)
from vllm.platforms import current_platform
from vllm.utils import FlexibleArgumentParser
from vllm.v1.spec_decode.ngram_proposer import NgramProposer
from vllm.v1.worker.gpu_input_batch import InputBatch
from vllm.v1.worker.gpu_model_runner import GPUModelRunner
def main(args):
def benchmark_propose(args):
rows = []
for max_ngram in args.max_ngram:
collector = TimeCollector(TimeCollector.US)
@ -69,10 +83,88 @@ def main(args):
)
def benchmark_batched_propose(args):
NUM_SPECULATIVE_TOKENS_NGRAM = 10
PROMPT_LOOKUP_MIN = 5
PROMPT_LOOKUP_MAX = 15
MAX_MODEL_LEN = int(1e7)
DEVICE = current_platform.device_type
model_config = ModelConfig(model="facebook/opt-125m", runner="generate")
speculative_config = SpeculativeConfig(
target_model_config=model_config,
target_parallel_config=ParallelConfig(),
method="ngram",
num_speculative_tokens=NUM_SPECULATIVE_TOKENS_NGRAM,
prompt_lookup_max=PROMPT_LOOKUP_MAX,
prompt_lookup_min=PROMPT_LOOKUP_MIN,
)
vllm_config = VllmConfig(
model_config=model_config,
cache_config=CacheConfig(),
speculative_config=speculative_config,
device_config=DeviceConfig(device=current_platform.device_type),
parallel_config=ParallelConfig(),
load_config=LoadConfig(),
scheduler_config=SchedulerConfig(),
)
# monkey patch vllm.v1.worker.gpu_model_runner.get_pp_group
mock_pp_group = mock.MagicMock()
mock_pp_group.world_size = 1
with mock.patch(
"vllm.v1.worker.gpu_model_runner.get_pp_group", return_value=mock_pp_group
):
runner = GPUModelRunner(vllm_config, DEVICE)
# hack max model len
runner.max_model_len = MAX_MODEL_LEN
runner.drafter.max_model_len = MAX_MODEL_LEN
dummy_input_batch = InputBatch(
max_num_reqs=args.num_req,
max_model_len=MAX_MODEL_LEN,
max_num_batched_tokens=args.num_req * args.num_token,
device=DEVICE,
pin_memory=False,
vocab_size=256000,
block_sizes=[16],
)
dummy_input_batch._req_ids = list(str(id) for id in range(args.num_req))
dummy_input_batch.spec_decode_unsupported_reqs = ()
dummy_input_batch.num_tokens_no_spec = [args.num_token] * args.num_req
dummy_input_batch.token_ids_cpu = np.random.randint(
0, 20, (args.num_req, args.num_token)
)
runner.input_batch = dummy_input_batch
sampled_token_ids = [[0]] * args.num_req
print("Starting benchmark")
# first run is warmup so ignore it
for _ in range(args.num_iteration):
start = time.time()
runner.drafter.propose(
sampled_token_ids,
dummy_input_batch.req_ids,
dummy_input_batch.num_tokens_no_spec,
dummy_input_batch.token_ids_cpu,
dummy_input_batch.spec_decode_unsupported_reqs,
)
end = time.time()
print(f"Iteration time (s): {end - start}")
def invoke_main() -> None:
parser = FlexibleArgumentParser(
description="Benchmark the performance of N-gram speculative decode drafting"
)
parser.add_argument(
"--batched", action="store_true", help="consider time to prepare batch"
) # noqa: E501
parser.add_argument(
"--num-iteration",
type=int,
@ -105,8 +197,17 @@ def invoke_main() -> None:
help="Number of speculative tokens to generate",
)
args = parser.parse_args()
main(args)
if not args.batched:
benchmark_propose(args)
else:
benchmark_batched_propose(args)
"""
# Example command lines:
# time python3 benchmarks/benchmark_ngram_proposer.py
# time python3 benchmarks/benchmark_ngram_proposer.py --batched --num-iteration 4 --num-token 1000000 --num-req 128
""" # noqa: E501
if __name__ == "__main__":
invoke_main() # pragma: no cover

View File

@ -449,7 +449,8 @@ async def benchmark(
def prepare_extra_body(request) -> dict:
extra_body = {}
# Add the schema to the extra_body
extra_body[request.structure_type] = request.schema
extra_body["structured_outputs"] = {}
extra_body["structured_outputs"][request.structure_type] = request.schema
return extra_body
print("Starting initial single prompt test run...")

View File

@ -17,7 +17,7 @@ from weight_shapes import WEIGHT_SHAPES
from vllm import _custom_ops as ops
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
w8a8_block_fp8_matmul,
w8a8_triton_block_scaled_mm,
)
from vllm.utils import FlexibleArgumentParser, cdiv
@ -158,7 +158,7 @@ def bench_fp8(
"cutlass_fp8_fp8_fp16_scaled_mm_bias": lambda: ops.cutlass_scaled_mm(
a, b, scale_a, scale_b, torch.float16, bias.to(dtype=torch.float16)
),
"triton_fp8_fp8_fp16_scaled_mm_blockwise": lambda: w8a8_block_fp8_matmul(
"triton_fp8_fp8_fp16_scaled_mm_blockwise": lambda: w8a8_triton_block_scaled_mm(
a_cont, b.t(), block_scale_a, block_scale_b.t(), (128, 128)
),
"cutlass_fp8_fp8_fp16_scaled_mm_blockwise": lambda: ops.cutlass_scaled_mm(

View File

@ -55,9 +55,7 @@ benchmark() {
output_len=$2
CUDA_VISIBLE_DEVICES=0 python3 \
-m vllm.entrypoints.openai.api_server \
--model $model \
CUDA_VISIBLE_DEVICES=0 vllm serve $model \
--port 8100 \
--max-model-len 10000 \
--gpu-memory-utilization 0.6 \
@ -65,9 +63,7 @@ benchmark() {
'{"kv_connector":"P2pNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
CUDA_VISIBLE_DEVICES=1 python3 \
-m vllm.entrypoints.openai.api_server \
--model $model \
CUDA_VISIBLE_DEVICES=1 vllm serve $model \
--port 8200 \
--max-model-len 10000 \
--gpu-memory-utilization 0.6 \

View File

@ -38,16 +38,12 @@ wait_for_server() {
launch_chunked_prefill() {
model="meta-llama/Meta-Llama-3.1-8B-Instruct"
# disagg prefill
CUDA_VISIBLE_DEVICES=0 python3 \
-m vllm.entrypoints.openai.api_server \
--model $model \
CUDA_VISIBLE_DEVICES=0 vllm serve $model \
--port 8100 \
--max-model-len 10000 \
--enable-chunked-prefill \
--gpu-memory-utilization 0.6 &
CUDA_VISIBLE_DEVICES=1 python3 \
-m vllm.entrypoints.openai.api_server \
--model $model \
CUDA_VISIBLE_DEVICES=1 vllm serve $model \
--port 8200 \
--max-model-len 10000 \
--enable-chunked-prefill \
@ -62,18 +58,14 @@ launch_chunked_prefill() {
launch_disagg_prefill() {
model="meta-llama/Meta-Llama-3.1-8B-Instruct"
# disagg prefill
CUDA_VISIBLE_DEVICES=0 python3 \
-m vllm.entrypoints.openai.api_server \
--model $model \
CUDA_VISIBLE_DEVICES=0 vllm serve $model \
--port 8100 \
--max-model-len 10000 \
--gpu-memory-utilization 0.6 \
--kv-transfer-config \
'{"kv_connector":"P2pNcclConnector","kv_role":"kv_producer","kv_rank":0,"kv_parallel_size":2,"kv_buffer_size":5e9}' &
CUDA_VISIBLE_DEVICES=1 python3 \
-m vllm.entrypoints.openai.api_server \
--model $model \
CUDA_VISIBLE_DEVICES=1 vllm serve $model \
--port 8200 \
--max-model-len 10000 \
--gpu-memory-utilization 0.6 \

View File

@ -3,6 +3,7 @@
import argparse
import copy
import itertools
import os
import torch
from weight_shapes import WEIGHT_SHAPES
@ -23,21 +24,45 @@ PROVIDER_CFGS = {
"torch-bf16": dict(enabled=True),
"nvfp4": dict(no_a_quant=False, enabled=True),
"nvfp4-noquant": dict(no_a_quant=True, enabled=True),
"fbgemm-nvfp4": dict(fbgemm=True, no_a_quant=False, enabled=True),
"fbgemm-nvfp4-noquant": dict(fbgemm=True, no_a_quant=True, enabled=True),
}
_needs_fbgemm = any(
v.get("fbgemm", False) for v in PROVIDER_CFGS.values() if v.get("enabled", False)
)
if _needs_fbgemm:
try:
from fbgemm_gpu.experimental.gemm.triton_gemm.fp4_quantize import (
triton_scale_nvfp4_quant,
)
except ImportError:
print(
"WARNING: FBGEMM providers are enabled but fbgemm_gpu is not installed. "
"These providers will be skipped. Please install fbgemm_gpu with: "
"'pip install fbgemm-gpu-genai' to run them."
)
# Disable FBGEMM providers so the benchmark can run.
for cfg in PROVIDER_CFGS.values():
if cfg.get("fbgemm"):
cfg["enabled"] = False
_enabled = [k for k, v in PROVIDER_CFGS.items() if v["enabled"]]
def _quant_weight_nvfp4(b: torch.Tensor, device: str):
def _quant_weight_nvfp4(b: torch.Tensor, device: str, cfg):
# Compute global scale for weight
b_amax = torch.abs(b).max().to(torch.float32)
b_global_scale = FLOAT8_E4M3_MAX * FLOAT4_E2M1_MAX / b_amax
b_fp4, scale_b_fp4 = ops.scaled_fp4_quant(b, b_global_scale)
if "fbgemm" in cfg and cfg["fbgemm"]:
b_fp4, scale_b_fp4 = triton_scale_nvfp4_quant(b, b_global_scale)
else:
b_fp4, scale_b_fp4 = ops.scaled_fp4_quant(b, b_global_scale)
return b_fp4, scale_b_fp4, b_global_scale
def build_nvfp4_runner(cfg, a, b, dtype, device):
b_fp4, scale_b_fp4, b_global_scale = _quant_weight_nvfp4(b, device)
b_fp4, scale_b_fp4, b_global_scale = _quant_weight_nvfp4(b, device, cfg)
# Compute global scale for activation
# NOTE: This is generally provided ahead-of-time by the model checkpoint.
@ -46,6 +71,35 @@ def build_nvfp4_runner(cfg, a, b, dtype, device):
# Alpha for the GEMM operation
alpha = 1.0 / (a_global_scale * b_global_scale)
if "fbgemm" in cfg and cfg["fbgemm"]:
if cfg["no_a_quant"]:
a_fp4, scale_a_fp4 = triton_scale_nvfp4_quant(a, a_global_scale)
def run():
return torch.ops.fbgemm.f4f4bf16(
a_fp4,
b_fp4,
scale_a_fp4,
scale_b_fp4,
global_scale=alpha,
use_mx=False,
)
return run
else:
def run():
a_fp4, scale_a_fp4 = triton_scale_nvfp4_quant(a, a_global_scale)
return torch.ops.fbgemm.f4f4bf16(
a_fp4,
b_fp4,
scale_a_fp4,
scale_b_fp4,
global_scale=alpha,
use_mx=False,
)
return run
if cfg["no_a_quant"]:
# Pre-quantize activation
@ -130,10 +184,13 @@ if __name__ == "__main__":
for K, N, model in prepare_shapes(args):
print(f"{model}, N={N} K={K}, BF16 vs NVFP4 GEMMs TFLOP/s:")
save_dir = f"bench_nvfp4_res_n{N}_k{K}"
os.makedirs(save_dir, exist_ok=True)
benchmark.run(
print_data=True,
show_plots=True,
save_path=f"bench_nvfp4_res_n{N}_k{K}",
save_path=save_dir,
N=N,
K=K,
)

View File

@ -51,7 +51,7 @@ def calculate_diff(
):
"""Calculate the difference between Inductor and CUDA implementations."""
device = torch.device("cuda")
x = torch.rand((batch_size * hidden_size, 4096), dtype=dtype, device=device)
x = torch.randn((batch_size, hidden_size), dtype=dtype, device=device)
quant_fp8 = QuantFP8(False, group_shape, column_major_scales=False)
@ -59,23 +59,25 @@ def calculate_diff(
torch_eager_out, torch_eager_scale = quant_fp8.forward_native(x)
cuda_out, cuda_scale = quant_fp8.forward_cuda(x)
out_allclose = lambda o1, o2: torch.allclose(
o1.to(torch.float32),
o2.to(torch.float32),
rtol=1e-3,
atol=1e-5,
)
scale_allclose = lambda s1, s2: torch.allclose(s1, s2, rtol=1e-3, atol=1e-5)
if (
out_allclose(cuda_out, torch_out)
and scale_allclose(cuda_scale, torch_scale)
and out_allclose(cuda_out, torch_eager_out)
and scale_allclose(cuda_scale, torch_eager_scale)
):
try:
torch.testing.assert_close(
cuda_out.to(torch.float32),
torch_out.to(torch.float32),
rtol=1e-3,
atol=1e-5,
)
torch.testing.assert_close(cuda_scale, torch_scale, rtol=1e-3, atol=1e-5)
torch.testing.assert_close(
cuda_out.to(torch.float32),
torch_eager_out.to(torch.float32),
rtol=1e-3,
atol=1e-5,
)
torch.testing.assert_close(cuda_scale, torch_eager_scale, rtol=1e-3, atol=1e-5)
print("✅ All implementations match")
else:
except AssertionError as e:
print("❌ Implementations differ")
print(e)
configs = []
@ -91,7 +93,7 @@ def benchmark_quantization(
):
device = torch.device("cuda")
x = torch.randn(batch_size * hidden_size, 4096, device=device, dtype=dtype)
x = torch.randn(batch_size, hidden_size, device=device, dtype=dtype)
quantiles = [0.5, 0.2, 0.8]
quant_fp8 = QuantFP8(False, group_shape, column_major_scales=col_major)
@ -157,21 +159,21 @@ if __name__ == "__main__":
)
parser.add_argument("-c", "--check", action="store_true")
parser.add_argument(
"--dtype", type=str, choices=["half", "bfloat16", "float"], default="half"
"--dtype", type=str, choices=["half", "bfloat16", "float"], default="bfloat16"
)
parser.add_argument(
"--hidden-sizes",
type=int,
nargs="+",
default=None,
help="Hidden sizes to benchmark (default: 1,16,64,128,256,512,1024,2048,4096)",
default=[896, 1024, 2048, 4096, 7168],
help="Hidden sizes to benchmark",
)
parser.add_argument(
"--batch-sizes",
type=int,
nargs="+",
default=None,
help="Batch sizes to benchmark (default: 1,16,32,64,128)",
default=[1, 16, 128, 512, 1024],
help="Batch sizes to benchmark",
)
parser.add_argument(
"--group-sizes",
@ -192,8 +194,8 @@ if __name__ == "__main__":
dtype = STR_DTYPE_TO_TORCH_DTYPE[args.dtype]
hidden_sizes = args.hidden_sizes or [1, 16, 64, 128, 256, 512, 1024, 2048, 4096]
batch_sizes = args.batch_sizes or [1, 16, 32, 64, 128]
hidden_sizes = args.hidden_sizes
batch_sizes = args.batch_sizes
if args.group_sizes is not None:
group_shapes = []

View File

@ -0,0 +1,406 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""
Benchmark the performance of the cutlass_moe_fp8 kernel vs the triton_moe
kernel. Both kernels take in fp8 quantized weights and 16-bit activations,
but use different quantization strategies and backends.
"""
import nvtx
import torch
from vllm import _custom_ops as ops
from vllm.model_executor.layers.fused_moe.config import fp8_w8a8_moe_quant_config
from vllm.model_executor.layers.fused_moe.cutlass_moe import cutlass_moe_fp8
from vllm.model_executor.layers.fused_moe.fused_moe import fused_experts, fused_topk
from vllm.platforms import current_platform
from vllm.utils import FlexibleArgumentParser
# Weight shapes for different models: [num_experts, topk, hidden_size,
# intermediate_size]
WEIGHT_SHAPES_MOE = {
"mixtral-8x7b": [
[8, 2, 4096, 14336],
],
"deepseek-v2": [
[160, 6, 5120, 12288],
],
"custom-small": [
[8, 2, 2048, 7168],
],
"glm45-fp8": [
[128, 8, 4096, 1408],
],
"Llama-4-Maverick-17B-128E-Instruct-FP8": [
[128, 1, 5120, 8192],
],
}
DEFAULT_MODELS = [
"mixtral-8x7b",
]
DEFAULT_BATCH_SIZES = [4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048]
DEFAULT_TP_SIZES = [1]
PER_ACT_TOKEN_OPTS = [False, True]
PER_OUT_CH_OPTS = [False, True]
FP8_DTYPE = current_platform.fp8_dtype()
def bench_run(
results: list,
model: str,
num_experts: int,
topk: int,
per_act_token: bool,
per_out_ch: bool,
mkn: tuple[int, int, int],
):
(m, k, n) = mkn
dtype = torch.half
device = "cuda"
# Create input activations
a = torch.randn((m, k), device=device, dtype=dtype) / 10
# Create weights
w1 = torch.randn((num_experts, 2 * n, k), device=device, dtype=dtype) / 10
w2 = torch.randn((num_experts, k, n), device=device, dtype=dtype) / 10
# Create FP8 quantized weights and scales for both kernels
w1_fp8q = torch.empty((num_experts, 2 * n, k), device=device, dtype=FP8_DTYPE)
w2_fp8q = torch.empty((num_experts, k, n), device=device, dtype=FP8_DTYPE)
# Create scales based on quantization strategy
if per_out_ch:
# Per-channel quantization
w1_scale = torch.empty(
(num_experts, 2 * n, 1), device=device, dtype=torch.float32
)
w2_scale = torch.empty((num_experts, k, 1), device=device, dtype=torch.float32)
else:
# Per-tensor quantization
w1_scale = torch.empty((num_experts, 1, 1), device=device, dtype=torch.float32)
w2_scale = torch.empty((num_experts, 1, 1), device=device, dtype=torch.float32)
# Quantize weights
for expert in range(num_experts):
if per_out_ch:
# Per-channel quantization - not yet implemented properly
# For now, fall back to per-tensor quantization
w1_fp8q[expert], w1_scale_temp = ops.scaled_fp8_quant(w1[expert])
w2_fp8q[expert], w2_scale_temp = ops.scaled_fp8_quant(w2[expert])
# Expand scalar scales to the expected per-channel shape
w1_scale[expert] = w1_scale_temp.expand(2 * n, 1)
w2_scale[expert] = w2_scale_temp.expand(k, 1)
else:
# Per-tensor quantization
w1_fp8q[expert], w1_scale_temp = ops.scaled_fp8_quant(w1[expert])
w2_fp8q[expert], w2_scale_temp = ops.scaled_fp8_quant(w2[expert])
# Store scalar scales in [1, 1] tensors
w1_scale[expert, 0, 0] = w1_scale_temp
w2_scale[expert, 0, 0] = w2_scale_temp
# Prepare weights for CUTLASS (no transpose needed)
w1_fp8q_cutlass = w1_fp8q # Keep original [E, 2N, K]
w2_fp8q_cutlass = w2_fp8q # Keep original [E, K, N]
# Create router scores and get topk
score = torch.randn((m, num_experts), device=device, dtype=dtype)
topk_weights, topk_ids, _ = fused_topk(a, score, topk, renormalize=False)
# WORKAROUND: CUTLASS MoE FP8 has issues with per-token quantization
# Force per-tensor quantization for all cases to match working e2e setup
a1_scale = torch.full((), 1e-2, device=device, dtype=torch.float32)
a2_scale = torch.full((), 1e-2, device=device, dtype=torch.float32)
# Force per-tensor quantization for all cases
per_act_token = False
# Create stride tensors for CUTLASS
ab_strides1 = torch.full((num_experts,), k, dtype=torch.int64, device=device)
ab_strides2 = torch.full((num_experts,), n, dtype=torch.int64, device=device)
c_strides1 = torch.full((num_experts,), 2 * n, dtype=torch.int64, device=device)
c_strides2 = torch.full((num_experts,), k, dtype=torch.int64, device=device)
def run_triton_moe(
a: torch.Tensor,
w1: torch.Tensor,
w2: torch.Tensor,
topk_weights: torch.Tensor,
topk_ids: torch.Tensor,
w1_scale: torch.Tensor,
w2_scale: torch.Tensor,
a1_scale: torch.Tensor,
a2_scale: torch.Tensor,
num_repeats: int,
):
quant_config = fp8_w8a8_moe_quant_config(
w1_scale=w1_scale,
w2_scale=w2_scale,
a1_scale=a1_scale,
a2_scale=a2_scale,
per_act_token_quant=per_act_token,
per_out_ch_quant=per_out_ch,
)
for _ in range(num_repeats):
fused_experts(
a,
w1,
w2,
topk_weights,
topk_ids,
quant_config=quant_config,
)
def run_cutlass_moe_fp8(
a: torch.Tensor,
w1: torch.Tensor,
w2: torch.Tensor,
topk_weights: torch.Tensor,
topk_ids: torch.Tensor,
ab_strides1: torch.Tensor,
ab_strides2: torch.Tensor,
c_strides1: torch.Tensor,
c_strides2: torch.Tensor,
w1_scale: torch.Tensor,
w2_scale: torch.Tensor,
a1_scale: torch.Tensor,
a2_scale: torch.Tensor,
num_repeats: int,
):
quant_config = fp8_w8a8_moe_quant_config(
w1_scale=w1_scale,
w2_scale=w2_scale,
a1_scale=a1_scale,
a2_scale=a2_scale,
per_act_token_quant=per_act_token,
per_out_ch_quant=per_out_ch,
)
for _ in range(num_repeats):
with nvtx.annotate("cutlass_moe_fp8", color="blue"):
cutlass_moe_fp8(
a=a,
w1_q=w1,
w2_q=w2,
topk_weights=topk_weights,
topk_ids=topk_ids,
ab_strides1=ab_strides1,
ab_strides2=ab_strides2,
c_strides1=c_strides1,
c_strides2=c_strides2,
quant_config=quant_config,
activation="silu",
global_num_experts=num_experts,
)
# Pre-create quantization config to avoid creating it inside CUDA graph
quant_config = fp8_w8a8_moe_quant_config(
w1_scale=w1_scale,
w2_scale=w2_scale,
a1_scale=a1_scale,
a2_scale=a2_scale,
per_act_token_quant=per_act_token,
per_out_ch_quant=per_out_ch,
)
# Create CUDA graphs for CUTLASS (match benchmark_moe.py pattern exactly)
cutlass_stream = torch.cuda.Stream()
cutlass_graph = torch.cuda.CUDAGraph()
with torch.cuda.graph(cutlass_graph, stream=cutlass_stream):
# Capture 10 invocations like benchmark_moe.py
for _ in range(10):
cutlass_moe_fp8(
a=a,
w1_q=w1_fp8q_cutlass,
w2_q=w2_fp8q_cutlass,
topk_weights=topk_weights,
topk_ids=topk_ids,
ab_strides1=ab_strides1,
ab_strides2=ab_strides2,
c_strides1=c_strides1,
c_strides2=c_strides2,
quant_config=quant_config,
activation="silu",
global_num_experts=num_experts,
)
torch.cuda.synchronize()
# Create CUDA graphs for Triton (match benchmark_moe.py pattern exactly)
triton_stream = torch.cuda.Stream()
triton_graph = torch.cuda.CUDAGraph()
with torch.cuda.graph(triton_graph, stream=triton_stream):
# Capture 10 invocations like benchmark_moe.py
for _ in range(10):
fused_experts(
a,
w1_fp8q,
w2_fp8q,
topk_weights,
topk_ids,
quant_config=quant_config,
)
torch.cuda.synchronize()
def bench_cuda_graph(graph, num_warmup=5, num_iters=100):
"""Benchmark CUDA graph using events like benchmark_moe.py"""
# Warmup
for _ in range(num_warmup):
graph.replay()
torch.cuda.synchronize()
# Timing
start_event = torch.cuda.Event(enable_timing=True)
end_event = torch.cuda.Event(enable_timing=True)
latencies = []
for _ in range(num_iters):
torch.cuda.synchronize()
start_event.record()
graph.replay()
end_event.record()
end_event.synchronize()
latencies.append(start_event.elapsed_time(end_event))
# Divide by 10 since graph contains 10 calls
return sum(latencies) / (num_iters * 10)
# Benchmark parameters
num_warmup = 5
num_iters = 100
# Benchmark only CUDA graphs (more reliable and faster)
# Benchmark Triton MoE with CUDA graphs
triton_graph_time = bench_cuda_graph(
triton_graph, num_warmup=num_warmup, num_iters=num_iters
)
# Benchmark CUTLASS MoE with CUDA graphs
cutlass_graph_time = bench_cuda_graph(
cutlass_graph, num_warmup=num_warmup, num_iters=num_iters
)
# Convert ms to us and return results
triton_time_us = triton_graph_time * 1000
cutlass_time_us = cutlass_graph_time * 1000
return {
"batch_size": m,
"triton_time_us": triton_time_us,
"cutlass_time_us": cutlass_time_us,
}
def main(args):
print("Benchmarking models:")
for i, model in enumerate(args.models):
print(f"[{i}] {model}")
all_results = []
for model in args.models:
for tp in args.tp_sizes:
for layer in WEIGHT_SHAPES_MOE[model]:
num_experts = layer[0]
topk = layer[1]
size_k = layer[2]
size_n = layer[3] // tp
if len(args.limit_k) > 0 and size_k not in args.limit_k:
continue
if len(args.limit_n) > 0 and size_n not in args.limit_n:
continue
for per_act_token in args.per_act_token_opts:
for per_out_ch in args.per_out_ch_opts:
print(
f"\n=== {model}, experts={num_experts}, topk={topk},"
f"per_act={per_act_token}, per_out_ch={per_out_ch} ==="
)
config_results = []
for size_m in args.batch_sizes:
mkn = (size_m, size_k, size_n)
result = bench_run(
[], # Not used anymore
model,
num_experts,
topk,
per_act_token,
per_out_ch,
mkn,
)
if result:
config_results.append(result)
# Print results table for this configuration
if config_results:
print(
f"\n{'Batch Size':<12}"
f"{'Triton (us)':<15}"
f"{'CUTLASS (us)':<15}"
)
print("-" * 45)
for result in config_results:
print(
f"{result['batch_size']:<12}"
f"{result['triton_time_us']:<15.2f}"
f"{result['cutlass_time_us']:<15.2f}"
)
all_results.extend(config_results)
print(f"\nTotal benchmarks completed: {len(all_results)}")
if __name__ == "__main__":
parser = FlexibleArgumentParser(
description="""Benchmark CUTLASS FP8 MOE vs Triton FP8 FUSED MOE
across specified models/shapes/batches
Example usage:
python benchmark_cutlass_moe_fp8.py \
--model "Llama-4-Maverick-17B-128E-Instruct-FP8" \
--tp-sizes 8 \
--batch-size 2 4 8 \
--per-act-token-opts false \
--per-out-ch-opts false
"""
)
parser.add_argument(
"--models",
nargs="+",
type=str,
default=DEFAULT_MODELS,
choices=WEIGHT_SHAPES_MOE.keys(),
)
parser.add_argument("--tp-sizes", nargs="+", type=int, default=DEFAULT_TP_SIZES)
parser.add_argument(
"--batch-sizes", nargs="+", type=int, default=DEFAULT_BATCH_SIZES
)
parser.add_argument("--limit-k", nargs="+", type=int, default=[])
parser.add_argument("--limit-n", nargs="+", type=int, default=[])
parser.add_argument(
"--per-act-token-opts",
nargs="+",
type=lambda x: x.lower() == "true",
default=[False, True],
help="Per-activation token quantization options (true/false)",
)
parser.add_argument(
"--per-out-ch-opts",
nargs="+",
type=lambda x: x.lower() == "true",
default=[False, True],
help="Per-output channel quantization options (true/false)",
)
args = parser.parse_args()
main(args)

View File

@ -7,6 +7,10 @@ Benchmark script for device communicators:
CustomAllreduce (oneshot, twoshot), PyNcclCommunicator,
and SymmMemCommunicator (multimem, two-shot).
for NCCL symmetric memory you need to set the environment variables
NCCL_NVLS_ENABLE=1 NCCL_CUMEM_ENABLE=1 VLLM_USE_NCCL_SYMM_MEM=1, otherwise NCCL does
not use fast NVLS implementation for all reduce.
Usage:
torchrun --nproc_per_node=<N> benchmark_device_communicators.py [options]
@ -26,7 +30,13 @@ import torch.distributed as dist
from torch.distributed import ProcessGroup
from vllm.distributed.device_communicators.custom_all_reduce import CustomAllreduce
from vllm.distributed.device_communicators.pynccl import PyNcclCommunicator
from vllm.distributed.device_communicators.pynccl import (
PyNcclCommunicator,
register_nccl_symmetric_ops,
)
from vllm.distributed.device_communicators.pynccl_allocator import (
set_graph_pool_id,
)
from vllm.distributed.device_communicators.symm_mem import SymmMemCommunicator
from vllm.logger import init_logger
from vllm.utils import FlexibleArgumentParser
@ -98,6 +108,7 @@ class CommunicatorBenchmark:
)
if not self.pynccl_comm.disabled:
logger.info("Rank %s: PyNcclCommunicator initialized", self.rank)
register_nccl_symmetric_ops(self.pynccl_comm)
else:
logger.info("Rank %s: PyNcclCommunicator disabled", self.rank)
self.pynccl_comm = None
@ -194,6 +205,15 @@ class CommunicatorBenchmark:
None, # no env variable needed
)
)
communicators.append(
(
"pynccl-symm",
lambda t: torch.ops.vllm.all_reduce_symmetric_with_copy(t),
lambda t: True, # Always available if initialized
nullcontext(),
None, # no env variable needed
)
)
if self.symm_mem_comm_multimem is not None:
comm = self.symm_mem_comm_multimem
@ -271,7 +291,9 @@ class CommunicatorBenchmark:
# Capture the graph using context manager
with context:
graph = torch.cuda.CUDAGraph()
with torch.cuda.graph(graph):
graph_pool = torch.cuda.graph_pool_handle()
set_graph_pool_id(graph_pool)
with torch.cuda.graph(graph, pool=graph_pool):
for _ in range(CUDA_GRAPH_CAPTURE_CYCLES):
allreduce_fn(graph_input)

View File

@ -79,9 +79,9 @@ def make_rand_lora_weight_tensor(
def make_rand_tensors(
a_shape: tuple[int],
b_shape: tuple[int],
c_shape: tuple[int],
a_shape: tuple[int, ...],
b_shape: tuple[int, ...],
c_shape: tuple[int, ...],
a_dtype: torch.dtype,
b_dtype: torch.dtype,
c_dtype: torch.dtype,
@ -243,7 +243,7 @@ class OpType(Enum):
lora_rank: int,
num_loras: int,
num_slices: int,
) -> tuple[tuple[int], tuple[int], tuple[int]]:
) -> tuple[tuple[int, ...], tuple[int, ...], tuple[int, ...]]:
"""
Given num_slices, return the shapes of the A, B, and C matrices
in A x B = C, for the op_type

View File

@ -584,8 +584,9 @@ def main(args: argparse.Namespace):
topk = config.num_experts_per_tok
intermediate_size = config.intermediate_size
elif config.architectures[0] in (
"DeepseekV3ForCausalLM",
"DeepseekV2ForCausalLM",
"DeepseekV3ForCausalLM",
"DeepseekV32ForCausalLM",
"Glm4MoeForCausalLM",
):
E = config.n_routed_experts

View File

@ -0,0 +1,174 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from __future__ import annotations
import random
import time
import torch
from tabulate import tabulate
from vllm import _custom_ops as ops
from vllm.logger import init_logger
from vllm.platforms import current_platform
from vllm.utils import (
STR_DTYPE_TO_TORCH_DTYPE,
FlexibleArgumentParser,
create_kv_caches_with_random,
)
logger = init_logger(__name__)
@torch.inference_mode()
def run_benchmark(
num_tokens: int,
num_heads: int,
head_size: int,
block_size: int,
num_blocks: int,
dtype: torch.dtype,
kv_cache_dtype: str,
num_iters: int,
benchmark_mode: str,
device: str = "cuda",
) -> float:
"""Return latency (seconds) for given num_tokens."""
if kv_cache_dtype == "fp8" and head_size % 16:
raise ValueError("fp8 kv-cache requires head_size to be a multiple of 16.")
current_platform.seed_everything(42)
torch.set_default_device(device)
# create random key / value tensors [T, H, D].
key = torch.randn(num_tokens, num_heads, head_size, dtype=dtype, device=device)
value = torch.randn_like(key)
# prepare the slot mapping.
# each token is assigned a unique slot in the KV-cache.
num_slots = block_size * num_blocks
if num_tokens > num_slots:
raise ValueError("num_tokens cannot exceed the total number of cache slots")
slot_mapping_lst = random.sample(range(num_slots), num_tokens)
slot_mapping = torch.tensor(slot_mapping_lst, dtype=torch.long, device=device)
key_caches, value_caches = create_kv_caches_with_random(
num_blocks,
block_size,
1, # num_layers
num_heads,
head_size,
kv_cache_dtype,
dtype,
device=device,
)
key_cache, value_cache = key_caches[0], value_caches[0]
# to free unused memory
del key_caches, value_caches
# compute per-kernel scaling factors for fp8 conversion (if used).
k_scale = (key.amax() / 64.0).to(torch.float32)
v_scale = (value.amax() / 64.0).to(torch.float32)
function_under_test = lambda: ops.reshape_and_cache(
key, # noqa: F821
value, # noqa: F821
key_cache, # noqa: F821
value_cache, # noqa: F821
slot_mapping, # noqa: F821
kv_cache_dtype,
k_scale,
v_scale,
)
if benchmark_mode == "cudagraph":
g = torch.cuda.CUDAGraph()
with torch.cuda.graph(g):
function_under_test()
torch.cuda.synchronize()
function_under_test = lambda: g.replay()
def run_cuda_benchmark(n_iters: int) -> float:
nonlocal key, value, key_cache, value_cache, slot_mapping
torch.cuda.synchronize()
start = time.perf_counter()
for _ in range(n_iters):
function_under_test()
torch.cuda.synchronize()
end = time.perf_counter()
return (end - start) / n_iters
# warm-up
run_cuda_benchmark(3)
lat = run_cuda_benchmark(num_iters)
# free tensors to mitigate OOM when sweeping
del key, value, key_cache, value_cache, slot_mapping
torch.cuda.empty_cache()
return lat
def main(args):
rows = []
for exp in range(1, 17):
n_tok = 2**exp
lat = run_benchmark(
num_tokens=n_tok,
num_heads=args.num_heads,
head_size=args.head_size,
block_size=args.block_size,
num_blocks=args.num_blocks,
dtype=STR_DTYPE_TO_TORCH_DTYPE[args.dtype],
kv_cache_dtype=args.kv_cache_dtype,
num_iters=args.iters,
benchmark_mode=args.mode,
device="cuda",
)
rows.append([n_tok, lat * 1e6]) # convert to microseconds
print(f"Benchmark results for implementation cuda (measuring with {args.mode}):")
print(tabulate(rows, headers=["num_tokens", "latency (µs)"], floatfmt=".3f"))
if __name__ == "__main__":
parser = FlexibleArgumentParser()
parser.add_argument("--num-heads", type=int, default=128)
parser.add_argument(
"--head-size",
type=int,
choices=[64, 80, 96, 112, 120, 128, 192, 256],
default=128,
)
parser.add_argument("--block-size", type=int, choices=[16, 32], default=16)
parser.add_argument("--num-blocks", type=int, default=128 * 128)
parser.add_argument(
"--dtype",
type=str,
choices=["half", "bfloat16", "float"],
default="bfloat16",
)
parser.add_argument(
"--kv-cache-dtype",
type=str,
choices=["auto", "fp8"],
default="auto",
)
parser.add_argument("--iters", type=int, default=200)
parser.add_argument(
"--mode",
type=str,
choices=["cudagraph", "no_graph"],
default="cudagraph",
)
args = parser.parse_args()
main(args)

View File

@ -9,6 +9,9 @@ import torch
from tabulate import tabulate
from vllm import _custom_ops as ops
from vllm.attention.ops.triton_reshape_and_cache_flash import (
triton_reshape_and_cache_flash,
)
from vllm.logger import init_logger
from vllm.platforms import current_platform
from vllm.utils import (
@ -31,6 +34,8 @@ def run_benchmark(
kv_cache_dtype: str,
kv_cache_layout: str,
num_iters: int,
implementation: str,
benchmark_mode: str,
device: str = "cuda",
) -> float:
"""Return latency (seconds) for given num_tokens."""
@ -38,6 +43,14 @@ def run_benchmark(
if kv_cache_dtype == "fp8" and head_size % 16:
raise ValueError("fp8 kv-cache requires head_size to be a multiple of 16.")
if implementation not in ("cuda", "triton"):
raise ValueError(
f"Unsupported implementation: {implementation}. "
"Only 'cuda' and 'triton' are supported."
)
if implementation == "triton" and kv_cache_layout == "HND":
return float("nan") # Triton does not support HND layout yet.
current_platform.seed_everything(42)
torch.set_default_device(device)
@ -65,27 +78,49 @@ def run_benchmark(
cache_layout=kv_cache_layout,
)
key_cache, value_cache = key_caches[0], value_caches[0]
# to free unused memory
del key_caches, value_caches
# compute per-kernel scaling factors for fp8 conversion (if used).
k_scale = (key.amax() / 64.0).to(torch.float32)
v_scale = (value.amax() / 64.0).to(torch.float32)
if implementation == "cuda":
function_under_test = lambda: ops.reshape_and_cache_flash(
key, # noqa: F821
value, # noqa: F821
key_cache, # noqa: F821
value_cache, # noqa: F821
slot_mapping, # noqa: F821
kv_cache_dtype,
k_scale,
v_scale,
)
else:
function_under_test = lambda: triton_reshape_and_cache_flash(
key, # noqa: F821
value, # noqa: F821
key_cache, # noqa: F821
value_cache, # noqa: F821
slot_mapping, # noqa: F821
kv_cache_dtype,
k_scale,
v_scale,
)
if benchmark_mode == "cudagraph":
g = torch.cuda.CUDAGraph()
with torch.cuda.graph(g):
function_under_test()
torch.cuda.synchronize()
function_under_test = lambda: g.replay()
def run_cuda_benchmark(n_iters: int) -> float:
nonlocal key, value, key_cache, value_cache, slot_mapping
torch.cuda.synchronize()
start = time.perf_counter()
for _ in range(n_iters):
ops.reshape_and_cache_flash(
key,
value,
key_cache,
value_cache,
slot_mapping,
kv_cache_dtype,
k_scale,
v_scale,
)
torch.cuda.synchronize()
function_under_test()
torch.cuda.synchronize()
end = time.perf_counter()
return (end - start) / n_iters
@ -116,10 +151,16 @@ def main(args):
kv_cache_dtype=args.kv_cache_dtype,
kv_cache_layout=layout,
num_iters=args.iters,
implementation=args.implementation,
benchmark_mode=args.mode,
device="cuda",
)
rows.append([n_tok, layout, f"{lat * 1e6:.3f}"])
print(
f"Benchmark results for implementation {args.implementation}"
f" (measuring with {args.mode}):"
)
print(tabulate(rows, headers=["num_tokens", "layout", "latency (µs)"]))
@ -151,6 +192,21 @@ if __name__ == "__main__":
)
parser.add_argument("--iters", type=int, default=100)
parser.add_argument(
"--implementation",
type=str,
choices=["cuda", "triton"],
default="cuda",
)
parser.add_argument(
"--mode",
type=str,
choices=["cudagraph", "no_graph"],
default="cudagraph",
)
args = parser.parse_args()
main(args)

View File

@ -11,13 +11,13 @@ from datetime import datetime
from typing import Any
import torch
import triton
from tqdm import tqdm
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
_w8a8_block_fp8_matmul,
)
from vllm.platforms import current_platform
from vllm.triton_utils import triton
from vllm.utils import FlexibleArgumentParser
mp.set_start_method("spawn", force=True)

View File

@ -8,12 +8,16 @@ import torch
from vllm import _custom_ops as ops
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
get_col_major_tma_aligned_tensor,
per_token_group_quant_fp8,
w8a8_block_fp8_matmul,
w8a8_triton_block_scaled_mm,
)
from vllm.triton_utils import triton
from vllm.utils.deep_gemm import calc_diff, fp8_gemm_nt, per_block_cast_to_fp8
from vllm.utils.deep_gemm import (
calc_diff,
fp8_gemm_nt,
get_col_major_tma_aligned_tensor,
per_block_cast_to_fp8,
)
def benchmark_shape(m: int,
@ -59,7 +63,7 @@ def benchmark_shape(m: int,
# === vLLM Triton Implementation ===
def vllm_triton_gemm():
return w8a8_block_fp8_matmul(A_vllm,
return w8a8_triton_block_scaled_mm(A_vllm,
B_vllm,
A_scale_vllm,
B_scale_vllm,

View File

@ -101,6 +101,7 @@ else()
find_isa(${CPUINFO} "asimd" ASIMD_FOUND) # Check for ARM NEON support
find_isa(${CPUINFO} "bf16" ARM_BF16_FOUND) # Check for ARM BF16 support
find_isa(${CPUINFO} "S390" S390_FOUND)
find_isa(${CPUINFO} "v" RVV_FOUND) # Check for RISC-V RVV support
endif()
if (AVX512_FOUND AND NOT AVX512_DISABLED)
@ -177,8 +178,14 @@ elseif (S390_FOUND)
"-mzvector"
"-march=native"
"-mtune=native")
elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "riscv64")
if(RVV_FOUND)
message(FAIL_ERROR "Can't support rvv now.")
else()
list(APPEND CXX_COMPILE_FLAGS "-march=rv64gc")
endif()
else()
message(FATAL_ERROR "vLLM CPU backend requires AVX512, AVX2, Power9+ ISA, S390X ISA or ARMv8 support.")
message(FATAL_ERROR "vLLM CPU backend requires AVX512, AVX2, Power9+ ISA, S390X ISA, ARMv8 or RISC-V support.")
endif()
#
@ -258,7 +265,8 @@ set(VLLM_EXT_SRC
"csrc/cpu/layernorm.cpp"
"csrc/cpu/mla_decode.cpp"
"csrc/cpu/pos_encoding.cpp"
"csrc/cpu/torch_bindings.cpp")
"csrc/cpu/torch_bindings.cpp"
"csrc/moe/dynamic_4bit_int_moe_cpu.cpp")
if (AVX512_FOUND AND NOT AVX512_DISABLED)
set(VLLM_EXT_SRC

View File

@ -18,8 +18,8 @@ if(FLASH_MLA_SRC_DIR)
else()
FetchContent_Declare(
flashmla
GIT_REPOSITORY https://github.com/vllm-project/FlashMLA.git
GIT_TAG a757314c04eedd166e329e846c820eb1bdd702de
GIT_REPOSITORY https://github.com/vllm-project/FlashMLA
GIT_TAG 5f65b85703c7ed75fda01e06495077caad207c3f
GIT_PROGRESS TRUE
CONFIGURE_COMMAND ""
BUILD_COMMAND ""
@ -33,23 +33,64 @@ message(STATUS "FlashMLA is available at ${flashmla_SOURCE_DIR}")
# The FlashMLA kernels only work on hopper and require CUDA 12.3 or later.
# Only build FlashMLA kernels if we are building for something compatible with
# sm90a
cuda_archs_loose_intersection(FLASH_MLA_ARCHS "9.0a" "${CUDA_ARCHS}")
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.3 AND FLASH_MLA_ARCHS)
set(SUPPORT_ARCHS)
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.3)
list(APPEND SUPPORT_ARCHS 9.0a)
endif()
if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.8)
list(APPEND SUPPORT_ARCHS 10.0a)
endif()
cuda_archs_loose_intersection(FLASH_MLA_ARCHS "${SUPPORT_ARCHS}" "${CUDA_ARCHS}")
if(FLASH_MLA_ARCHS)
set(VLLM_FLASHMLA_GPU_FLAGS ${VLLM_GPU_FLAGS})
list(APPEND VLLM_FLASHMLA_GPU_FLAGS "--expt-relaxed-constexpr" "--expt-extended-lambda" "--use_fast_math")
set(FlashMLA_SOURCES
${flashmla_SOURCE_DIR}/csrc/flash_api.cpp
${flashmla_SOURCE_DIR}/csrc/kernels/get_mla_metadata.cu
${flashmla_SOURCE_DIR}/csrc/kernels/mla_combine.cu
${flashmla_SOURCE_DIR}/csrc/kernels/splitkv_mla.cu
${flashmla_SOURCE_DIR}/csrc/kernels_fp8/flash_fwd_mla_fp8_sm90.cu)
${flashmla_SOURCE_DIR}/csrc/torch_api.cpp
${flashmla_SOURCE_DIR}/csrc/pybind.cpp
${flashmla_SOURCE_DIR}/csrc/smxx/get_mla_metadata.cu
${flashmla_SOURCE_DIR}/csrc/smxx/mla_combine.cu
${flashmla_SOURCE_DIR}/csrc/sm90/decode/dense/splitkv_mla.cu
${flashmla_SOURCE_DIR}/csrc/sm90/decode/sparse_fp8/splitkv_mla.cu
${flashmla_SOURCE_DIR}/csrc/sm90/prefill/sparse/fwd.cu
${flashmla_SOURCE_DIR}/csrc/sm100/decode/sparse_fp8/splitkv_mla.cu
${flashmla_SOURCE_DIR}/csrc/sm100/prefill/dense/fmha_cutlass_fwd_sm100.cu
${flashmla_SOURCE_DIR}/csrc/sm100/prefill/dense/fmha_cutlass_bwd_sm100.cu
${flashmla_SOURCE_DIR}/csrc/sm100/prefill/sparse/fwd.cu
)
set(FlashMLA_Extension_SOURCES
${flashmla_SOURCE_DIR}/csrc/extension/torch_api.cpp
${flashmla_SOURCE_DIR}/csrc/extension/sm90/dense_fp8/pybind.cpp
${flashmla_SOURCE_DIR}/csrc/extension/sm90/dense_fp8/flash_fwd_mla_fp8_sm90.cu
)
set(FlashMLA_INCLUDES
${flashmla_SOURCE_DIR}/csrc
${flashmla_SOURCE_DIR}/csrc/sm90
${flashmla_SOURCE_DIR}/csrc/cutlass/include
${flashmla_SOURCE_DIR}/csrc)
${flashmla_SOURCE_DIR}/csrc/cutlass/tools/util/include
)
set(FlashMLA_Extension_INCLUDES
${flashmla_SOURCE_DIR}/csrc
${flashmla_SOURCE_DIR}/csrc/sm90
${flashmla_SOURCE_DIR}/csrc/extension/sm90/dense_fp8/
${flashmla_SOURCE_DIR}/csrc/cutlass/include
${flashmla_SOURCE_DIR}/csrc/cutlass/tools/util/include
)
set_gencode_flags_for_srcs(
SRCS "${FlashMLA_SOURCES}"
CUDA_ARCHS "${FLASH_MLA_ARCHS}")
set_gencode_flags_for_srcs(
SRCS "${FlashMLA_Extension_SOURCES}"
CUDA_ARCHS "${FLASH_MLA_ARCHS}")
define_gpu_extension_target(
_flashmla_C
DESTINATION vllm
@ -60,8 +101,32 @@ if(${CMAKE_CUDA_COMPILER_VERSION} VERSION_GREATER 12.3 AND FLASH_MLA_ARCHS)
INCLUDE_DIRECTORIES ${FlashMLA_INCLUDES}
USE_SABI 3
WITH_SOABI)
# Keep Stable ABI for the module, but *not* for CUDA/C++ files.
# This prevents Py_LIMITED_API from affecting nvcc and C++ compiles.
target_compile_options(_flashmla_C PRIVATE
$<$<COMPILE_LANGUAGE:CUDA>:-UPy_LIMITED_API>
$<$<COMPILE_LANGUAGE:CXX>:-UPy_LIMITED_API>)
define_gpu_extension_target(
_flashmla_extension_C
DESTINATION vllm
LANGUAGE ${VLLM_GPU_LANG}
SOURCES ${FlashMLA_Extension_SOURCES}
COMPILE_FLAGS ${VLLM_FLASHMLA_GPU_FLAGS}
ARCHITECTURES ${VLLM_GPU_ARCHES}
INCLUDE_DIRECTORIES ${FlashMLA_Extension_INCLUDES}
USE_SABI 3
WITH_SOABI)
# Keep Stable ABI for the module, but *not* for CUDA/C++ files.
# This prevents Py_LIMITED_API from affecting nvcc and C++ compiles.
target_compile_options(_flashmla_extension_C PRIVATE
$<$<COMPILE_LANGUAGE:CUDA>:-UPy_LIMITED_API>
$<$<COMPILE_LANGUAGE:CXX>:-UPy_LIMITED_API>)
else()
# Create an empty target for setup.py when not targeting sm90a systems
# Create empty targets for setup.py when not targeting sm90a systems
add_custom_target(_flashmla_C)
add_custom_target(_flashmla_extension_C)
endif()

View File

@ -38,7 +38,7 @@ else()
FetchContent_Declare(
vllm-flash-attn
GIT_REPOSITORY https://github.com/vllm-project/flash-attention.git
GIT_TAG ee4d25bd84e0cbc7e0b9b9685085fd5db2dcb62a
GIT_TAG 4695e6bed5366c41e28c06cd86170166e4f43d00
GIT_PROGRESS TRUE
# Don't share the vllm-flash-attn build between build types
BINARY_DIR ${CMAKE_BINARY_DIR}/vllm-flash-attn

View File

@ -310,13 +310,13 @@ function(cuda_archs_loose_intersection OUT_CUDA_ARCHS SRC_CUDA_ARCHS TGT_CUDA_AR
list(REMOVE_DUPLICATES _PTX_ARCHS)
list(REMOVE_DUPLICATES _SRC_CUDA_ARCHS)
# if x.0a is in SRC_CUDA_ARCHS and x.0 is in CUDA_ARCHS then we should
# remove x.0a from SRC_CUDA_ARCHS and add x.0a to _CUDA_ARCHS
# If x.0a or x.0f is in SRC_CUDA_ARCHS and x.0 is in CUDA_ARCHS then we should
# remove x.0a or x.0f from SRC_CUDA_ARCHS and add x.0a or x.0f to _CUDA_ARCHS
set(_CUDA_ARCHS)
foreach(_arch ${_SRC_CUDA_ARCHS})
if(_arch MATCHES "\\a$")
if(_arch MATCHES "[af]$")
list(REMOVE_ITEM _SRC_CUDA_ARCHS "${_arch}")
string(REPLACE "a" "" _base "${_arch}")
string(REGEX REPLACE "[af]$" "" _base "${_arch}")
if ("${_base}" IN_LIST TGT_CUDA_ARCHS)
list(REMOVE_ITEM _TGT_CUDA_ARCHS "${_base}")
list(APPEND _CUDA_ARCHS "${_arch}")

View File

@ -135,10 +135,10 @@ public:
max_splits = min(16, max_splits);
// TODO: This avoids a hang when the batch size larger than 1 and
// there is more than 4 kv_splits.
// there is more than 1 kv_splits.
// Discuss with NVIDIA how this can be fixed.
if (B > 1) {
max_splits = min(2, max_splits);
max_splits = min(1, max_splits);
}
// printf(" max_splits = %d\n", max_splits);

View File

@ -580,22 +580,22 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
for (; tile_scheduler.is_valid(); ++tile_scheduler) {
auto blk_coord = tile_scheduler.get_block_coord();
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
auto local_split_kv = params.split_kv;
if (params.mainloop.ptr_seq != nullptr) {
get<1>(problem_shape) = params.mainloop.ptr_seq[get<2>(blk_coord)];
if (params.ptr_split_kv != nullptr) {
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
}
if (local_split_kv <= get<3>(blk_coord))
continue;
if (local_split_kv <= get<3>(blk_coord))
continue;
load_page_table(
blk_coord,
problem_shape,
params.mainloop,
shared_storage.tensors,
pipeline_page_table, pipeline_pt_producer_state,
local_split_kv
local_split_kv
);
}
}
@ -604,15 +604,15 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
CUTLASS_PRAGMA_NO_UNROLL
for (; tile_scheduler.is_valid(); ++tile_scheduler) {
auto blk_coord = tile_scheduler.get_block_coord();
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
if (params.mainloop.ptr_seq != nullptr) {
get<1>(problem_shape) = params.mainloop.ptr_seq[get<2>(blk_coord)];
if (params.ptr_split_kv != nullptr) {
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
}
if (local_split_kv <= get<3>(blk_coord))
if (local_split_kv <= get<3>(blk_coord))
continue;
load_cpasync(
blk_coord,
@ -621,7 +621,7 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
params.mainloop_params,
shared_storage.tensors,
pipeline_load_qk, pipeline_load_qk_producer_state,
local_split_kv,
local_split_kv,
/* must be shared pipe */
pipeline_page_table, pipeline_pt_consumer_state
);
@ -633,15 +633,15 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
CUTLASS_PRAGMA_NO_UNROLL
for (; tile_scheduler.is_valid(); ++tile_scheduler) {
auto blk_coord = tile_scheduler.get_block_coord();
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
if (params.mainloop.ptr_seq != nullptr) {
get<1>(problem_shape) = params.mainloop.ptr_seq[get<2>(blk_coord)];
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
}
if (local_split_kv <= get<3>(blk_coord))
if (local_split_kv <= get<3>(blk_coord))
continue;
load_tma</* paged= */ true>(
blk_coord,
@ -651,7 +651,7 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
shared_storage.tensors,
pipeline_load_qk, pipeline_load_qk_producer_state,
pipeline_load_qk, pipeline_load_qk_producer_state,
local_split_kv
local_split_kv
);
cutlass::arch::NamedBarrier((kNumComputeWarps + kNumLoadWarps) * NumThreadsPerWarp, kNamedBarrierEpilogue).arrive_and_wait();
}
@ -660,15 +660,15 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
CUTLASS_PRAGMA_NO_UNROLL
for (; tile_scheduler.is_valid(); ++tile_scheduler) {
auto blk_coord = tile_scheduler.get_block_coord();
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
if (params.mainloop.ptr_seq != nullptr) {
get<1>(problem_shape) = params.mainloop.ptr_seq[get<2>(blk_coord)];
if (params.ptr_split_kv != nullptr) {
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
}
}
if (local_split_kv <= get<3>(blk_coord))
if (local_split_kv <= get<3>(blk_coord))
continue;
load_tma<false>(
blk_coord,
@ -678,7 +678,7 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
shared_storage.tensors,
pipeline_load_qk, pipeline_load_qk_producer_state,
pipeline_load_qk, pipeline_load_qk_producer_state,
local_split_kv
local_split_kv
);
cutlass::arch::NamedBarrier((kNumComputeWarps + kNumLoadWarps) * NumThreadsPerWarp, kNamedBarrierEpilogue).arrive_and_wait();
}
@ -694,14 +694,14 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
for (; tile_scheduler.is_valid(); ++tile_scheduler) {
auto blk_coord = tile_scheduler.get_block_coord();
auto problem_shape = params.problem_shape;
auto local_split_kv = params.split_kv;
auto local_split_kv = params.split_kv;
if (params.mainloop.ptr_seq != nullptr) {
get<1>(problem_shape) = params.mainloop.ptr_seq[get<2>(blk_coord)];
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
}
if (local_split_kv <= get<3>(blk_coord))
if (local_split_kv <= get<3>(blk_coord))
continue;
mma(blk_coord,
problem_shape,
@ -711,7 +711,7 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
pipeline_mma_s, pipeline_mma_s_producer_state,
pipeline_p_mma, pipeline_p_mma_consumer_state,
pipeline_mma_o, pipeline_mma_o_producer_state,
local_split_kv
local_split_kv
);
}
}
@ -726,15 +726,15 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
for (; tile_scheduler.is_valid(); ++tile_scheduler) {
auto blk_coord = tile_scheduler.get_block_coord();
auto problem_shape = params.problem_shape;
auto split_kv = params.split_kv;
auto local_split_kv = split_kv;
auto split_kv = params.split_kv;
auto local_split_kv = split_kv;
if (params.mainloop.ptr_seq != nullptr) {
get<1>(problem_shape) = params.mainloop.ptr_seq[get<2>(blk_coord)];
if (params.ptr_split_kv != nullptr) {
if (params.ptr_split_kv != nullptr) {
local_split_kv = params.ptr_split_kv[get<2>(blk_coord)];
}
}
if (local_split_kv <= get<3>(blk_coord))
if (local_split_kv <= get<3>(blk_coord))
continue;
compute(
blk_coord,
@ -745,7 +745,7 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
pipeline_mma_s, pipeline_mma_s_consumer_state,
pipeline_p_mma, pipeline_p_mma_producer_state,
pipeline_mma_o, pipeline_mma_o_consumer_state,
local_split_kv
local_split_kv
);
}
@ -1900,7 +1900,7 @@ struct Sm100FmhaMlaKernelTmaWarpspecialized {
cutlass::arch::NamedBarrier(
(kNumComputeWarps + kNumLoadWarps) * NumThreadsPerWarp,
kNamedBarrierEpilogue
).arrive();
).arrive_and_wait();
return;
}

View File

@ -56,3 +56,11 @@ void cp_gather_cache(
torch::Tensor const& block_table, // [BATCH, BLOCK_INDICES]
torch::Tensor const& cu_seq_lens, // [BATCH+1]
int64_t batch_size, std::optional<torch::Tensor> seq_starts = std::nullopt);
// Indexer K quantization and cache function
void indexer_k_quant_and_cache(
torch::Tensor& k, // [num_tokens, head_dim]
torch::Tensor& kv_cache, // [num_blocks, block_size, cache_stride]
torch::Tensor& slot_mapping, // [num_tokens]
int64_t quant_block_size, // quantization block size
const std::string& scale_fmt);

View File

@ -16,8 +16,7 @@
#include <algorithm>
#include <cassert>
#include <map>
#include <vector>
#include <cfloat> // FLT_MIN
#ifdef USE_ROCM
#include <hip/hip_bf16.h>
@ -209,6 +208,20 @@ void copy_blocks_mla(std::vector<torch::Tensor> const& kv_caches,
namespace vllm {
// Used to copy/convert one element
template <typename OutT, typename InT, Fp8KVCacheDataType kv_dt>
struct CopyWithScaleOp {
float scale;
__device__ __forceinline__ void operator()(OutT& dst, const InT src) const {
if constexpr (kv_dt == Fp8KVCacheDataType::kAuto) {
dst = static_cast<OutT>(src);
} else {
dst = fp8::scaled_convert<OutT, InT, kv_dt>(src, scale);
}
}
};
template <typename scalar_t, typename cache_t, Fp8KVCacheDataType kv_dt>
__global__ void reshape_and_cache_kernel(
const scalar_t* __restrict__ key, // [num_tokens, num_heads, head_size]
@ -224,59 +237,51 @@ __global__ void reshape_and_cache_kernel(
const int64_t token_idx = blockIdx.x;
const int64_t slot_idx = slot_mapping[token_idx];
if (slot_idx < 0) {
// Padding token that should be ignored.
return;
}
const int64_t block_idx = slot_idx / block_size;
const int64_t block_offset = slot_idx % block_size;
const int h_block_count = head_size / x; // head_size//x
const int n = num_heads * head_size;
for (int i = threadIdx.x; i < n; i += blockDim.x) {
const int64_t src_key_idx = token_idx * key_stride + i;
const int64_t src_value_idx = token_idx * value_stride + i;
const int h_block_idx = threadIdx.x;
if (h_block_idx >= num_heads * h_block_count) {
return;
}
const int head_idx = i / head_size;
const int head_offset = i % head_size;
const int x_idx = head_offset / x;
const int x_offset = head_offset % x;
const int head_idx = h_block_idx / h_block_count;
const int h_block = h_block_idx % h_block_count;
const int64_t tgt_key_idx =
block_idx * num_heads * (head_size / x) * block_size * x +
head_idx * (head_size / x) * block_size * x + x_idx * block_size * x +
block_offset * x + x_offset;
const int64_t tgt_value_idx =
block_idx * num_heads * head_size * block_size +
head_idx * head_size * block_size + head_offset * block_size +
block_offset;
scalar_t tgt_key = key[src_key_idx];
scalar_t tgt_value = value[src_value_idx];
if constexpr (kv_dt == Fp8KVCacheDataType::kAuto) {
key_cache[tgt_key_idx] = tgt_key;
value_cache[tgt_value_idx] = tgt_value;
} else {
key_cache[tgt_key_idx] =
fp8::scaled_convert<cache_t, scalar_t, kv_dt>(tgt_key, *k_scale);
value_cache[tgt_value_idx] =
fp8::scaled_convert<cache_t, scalar_t, kv_dt>(tgt_value, *v_scale);
}
const scalar_t* __restrict__ key_src =
key + token_idx * key_stride + head_idx * head_size + h_block * x;
const int64_t src_value_start =
token_idx * value_stride + head_idx * head_size + h_block * x;
cache_t* __restrict__ key_dst =
key_cache + block_idx * num_heads * h_block_count * block_size * x +
head_idx * h_block_count * block_size * x + h_block * block_size * x +
block_offset * x;
const int64_t tgt_value_start =
block_idx * num_heads * h_block_count * x * block_size +
head_idx * h_block_count * x * block_size + h_block * x * block_size +
block_offset;
constexpr int VEC_SIZE = (sizeof(scalar_t) == 2) ? 8 : 4;
float k_scale_val = (kv_dt == Fp8KVCacheDataType::kAuto) ? 0.f : *k_scale;
CopyWithScaleOp<cache_t, scalar_t, kv_dt> k_op{k_scale_val};
float v_scale_val = (kv_dt == Fp8KVCacheDataType::kAuto) ? 0.f : *v_scale;
CopyWithScaleOp<cache_t, scalar_t, kv_dt> v_op{v_scale_val};
vectorize_with_alignment<VEC_SIZE>(key_src, key_dst, x, 0, 1, k_op);
const scalar_t* __restrict__ value_src = value + src_value_start;
cache_t* __restrict__ value_dst = value_cache + tgt_value_start;
#pragma unroll
for (int i = 0; i < x; i++) {
v_op(value_dst[i * block_size], value_src[i]);
}
}
// Used by vectorization_utils to copy/convert one element
template <typename OutT, typename InT, Fp8KVCacheDataType kv_dt>
struct CopyWithScaleOp {
float scale;
__device__ __forceinline__ void operator()(OutT& dst, const InT src) const {
if constexpr (kv_dt == Fp8KVCacheDataType::kAuto) {
dst = static_cast<OutT>(src);
} else {
dst = fp8::scaled_convert<OutT, InT, kv_dt>(src, scale);
}
}
};
template <typename scalar_t, typename cache_t, Fp8KVCacheDataType kv_dt>
__global__ void reshape_and_cache_flash_kernel(
const scalar_t* __restrict__ key, // [num_tokens, num_heads, head_size]
@ -396,6 +401,176 @@ __global__ void concat_and_cache_mla_kernel(
copy(k_pe, kv_cache, k_pe_stride, block_stride, pe_dim, kv_lora_rank);
}
template <typename scalar_t, typename cache_t, Fp8KVCacheDataType kv_dt>
__global__ void concat_and_cache_ds_mla_kernel(
const scalar_t* __restrict__ kv_c, // [num_tokens, kv_lora_rank]
const scalar_t* __restrict__ k_pe, // [num_tokens, pe_dim]
cache_t* __restrict__ kv_cache, // [num_blocks, block_size, (kv_lora_rank
// + pe_dim)]
const int64_t* __restrict__ slot_mapping, // [num_tokens]
const int block_stride, //
const int entry_stride, //
const int kv_c_stride, //
const int k_pe_stride, //
const int kv_lora_rank, //
const int pe_dim, //
const int block_size, //
const float* scale //
) {
const int64_t token_idx = blockIdx.x;
const int64_t slot_idx = slot_mapping[token_idx];
// NOTE: slot_idx can be -1 if the token is padded
if (slot_idx < 0) {
return;
}
const int64_t block_idx = slot_idx / block_size;
const int64_t block_offset = slot_idx % block_size;
const int64_t dst_idx_start =
block_idx * block_stride + block_offset * entry_stride;
// For the NoPE part, each tile of 128 elements is handled by half of one warp
// (16 threads). There are 4 total tiles, so 2 warps (64 threads).
// Lanes 0 and 16 of each warp write the scale values for that warp's tiles.
// The RoPE part (last 64 elements) is handled by another 1 warp (32 threads).
// So in total, we use 3 warps (96 threads) per block.
// Cast kv_cache to 16_bit for RoPE values
scalar_t* kv_cache_16bit =
reinterpret_cast<scalar_t*>(&kv_cache[dst_idx_start]);
// The last warp handles the RoPE part
if (threadIdx.x >= 64) {
// Each thread handles two elements of RoPE
const int8_t pe_idx_start = (threadIdx.x - 64) * 2;
const int64_t src_idx = token_idx * k_pe_stride + pe_idx_start;
// Vectorized load of two 16-bit values, performed as one 32-bit load
const int32_t vals = *reinterpret_cast<const int32_t*>(&k_pe[src_idx]);
// RoPE values start after the packed 8-bit NoPE values and the
// 32-bit scales
const int64_t dst_idx = kv_lora_rank / 2 + 8 + pe_idx_start;
// Vectorized store of two 16-bit values, performed as one 32-bit store
*reinterpret_cast<int32_t*>(&kv_cache_16bit[dst_idx]) = vals;
return;
}
// The first two warps handle the NoPE part
const int8_t warp_idx = threadIdx.x >> 5;
const int8_t lane_idx = threadIdx.x & 31;
const int8_t tile_idx = warp_idx * 2 + (lane_idx >> 4);
// Each thread handles 8 elements of NoPE
// Load the NoPE elements for this thread into registers
const int64_t src_idx_start = token_idx * kv_c_stride + (threadIdx.x * 8);
// Vectorized load of eight 16-bit values, performed as an int4 load
const int4 vals_i4 = *reinterpret_cast<const int4*>(&kv_c[src_idx_start]);
const scalar_t* vals = reinterpret_cast<const scalar_t*>(&vals_i4);
// Max absolute value of this thread's elements
float max_abs = fmaxf(fmaxf(fmaxf(fabsf(vals[0]), fabsf(vals[1])),
fmaxf(fabsf(vals[2]), fabsf(vals[3]))),
fmaxf(fmaxf(fabsf(vals[4]), fabsf(vals[5])),
fmaxf(fabsf(vals[6]), fabsf(vals[7]))));
// Warp-level reduction to find the max absolute value in each half-warp
#pragma unroll
for (int offset = 8; offset > 0; offset /= 2) {
max_abs = fmaxf(max_abs, VLLM_SHFL_XOR_SYNC_WIDTH(max_abs, offset, 16));
}
// Compute the scale for the tile
float tile_scale = max_abs / 448.f;
// The first lane of each half-warp writes the scale to kv_cache
if ((lane_idx == 0) || (lane_idx == 16)) {
float* kv_cache_32bit = reinterpret_cast<float*>(&kv_cache[dst_idx_start]);
const uint64_t dst_idx = kv_lora_rank / 4 + tile_idx;
kv_cache_32bit[dst_idx] = tile_scale;
}
// Now all threads in the block scale and write their elements
// NoPE data is packed in the first kv_lora_rank/2 bytes (first 256 bytes)
const int64_t dst_idx_base = dst_idx_start + (threadIdx.x * 8);
uint8_t result[8];
#pragma unroll
for (int i = 0; i < 8; i++) {
result[i] =
fp8::scaled_convert<uint8_t, scalar_t, Fp8KVCacheDataType::kFp8E4M3>(
vals[i], tile_scale);
}
// Store as aligned 64-bit writes
*reinterpret_cast<uint64_t*>(&kv_cache[dst_idx_base]) =
*reinterpret_cast<const uint64_t*>(result);
}
template <typename scalar_t, typename cache_t, Fp8KVCacheDataType kv_dt>
__global__ void indexer_k_quant_and_cache_kernel(
const scalar_t* __restrict__ k, // [num_tokens, head_dim]
cache_t* __restrict__ kv_cache, // [num_blocks, block_size, cache_stride]
const int64_t* __restrict__ slot_mapping, // [num_tokens]
const int head_dim, // dimension of each head
const int quant_block_size, // quantization block size
const int cache_block_size, // cache block size
const int cache_stride, // stride for each token in kv_cache
const bool use_ue8m0 // use ue8m0 scale format
) {
constexpr int VEC_SIZE = 4;
const int64_t token_idx = blockIdx.x;
const int64_t head_dim_idx = (blockIdx.y * blockDim.y * blockDim.x +
threadIdx.y * blockDim.x + threadIdx.x) *
VEC_SIZE;
const int64_t slot_idx = slot_mapping[token_idx];
const int64_t block_idx = slot_idx / cache_block_size;
const int64_t block_offset = slot_idx % cache_block_size;
// NOTE: slot_idx can be -1 if the token is padded
if (slot_idx < 0 || (head_dim_idx >= head_dim)) {
return;
}
float2 k_val = (reinterpret_cast<const float2*>(
k))[(token_idx * head_dim + head_dim_idx) / VEC_SIZE];
scalar_t* k_val_ptr = reinterpret_cast<scalar_t*>(&k_val);
float amax = 0.0f;
for (int i = 0; i < VEC_SIZE; i++) {
amax = fmaxf(amax, fabsf(float(k_val_ptr[i])));
}
#ifndef USE_ROCM
__syncwarp();
#endif
// Reduced amax
for (int mask = 16; mask > 0; mask /= 2) {
#ifdef USE_ROCM
amax = fmaxf(amax, __shfl_xor_sync(uint64_t(-1), amax, mask));
#else
amax = fmaxf(amax, __shfl_xor_sync(unsigned(-1), amax, mask));
#endif
}
#ifndef USE_ROCM
__syncwarp();
#endif
float scale = fmaxf(amax, 1e-4) / 448.0f;
if (use_ue8m0) {
scale = exp2f(ceilf(log2f(scale)));
}
const int64_t dst_offset = block_idx * cache_block_size * cache_stride +
block_offset * head_dim + head_dim_idx;
for (int i = 0; i < VEC_SIZE; i++) {
kv_cache[dst_offset + i] =
fp8::scaled_convert<cache_t, scalar_t, kv_dt>(k_val_ptr[i], scale);
}
if (threadIdx.x == 0) {
const int64_t dst_scale_idx =
block_idx * cache_block_size * cache_stride +
cache_block_size * head_dim +
(block_offset * head_dim + head_dim_idx) * 4 / quant_block_size;
reinterpret_cast<float*>(kv_cache)[dst_scale_idx / 4] = scale;
}
}
} // namespace vllm
// KV_T is the data type of key and value tensors.
@ -431,14 +606,15 @@ void reshape_and_cache(
int key_stride = key.stride(0);
int value_stride = value.stride(0);
int head_div_x = head_size / x;
dim3 grid(num_tokens);
dim3 block(std::min(num_heads * head_size, 512));
dim3 block(std::min(num_heads * head_div_x, 512));
const at::cuda::OptionalCUDAGuard device_guard(device_of(key));
const cudaStream_t stream = at::cuda::getCurrentCUDAStream();
DISPATCH_BY_KV_CACHE_DTYPE(key.dtype(), kv_cache_dtype,
CALL_RESHAPE_AND_CACHE)
CALL_RESHAPE_AND_CACHE);
}
// KV_T is the data type of key and value tensors.
@ -509,6 +685,18 @@ void reshape_and_cache_flash(
kv_c_stride, k_pe_stride, kv_lora_rank, pe_dim, block_size, \
reinterpret_cast<const float*>(scale.data_ptr()));
// KV_T is the data type of key and value tensors.
// CACHE_T is the stored data type of kv-cache.
#define CALL_CONCAT_AND_CACHE_DS_MLA(KV_T, CACHE_T, KV_DTYPE) \
vllm::concat_and_cache_ds_mla_kernel<KV_T, CACHE_T, KV_DTYPE> \
<<<grid, block, 0, stream>>>( \
reinterpret_cast<KV_T*>(kv_c.data_ptr()), \
reinterpret_cast<KV_T*>(k_pe.data_ptr()), \
reinterpret_cast<CACHE_T*>(kv_cache.data_ptr()), \
slot_mapping.data_ptr<int64_t>(), block_stride, entry_stride, \
kv_c_stride, k_pe_stride, kv_lora_rank, pe_dim, block_size, \
reinterpret_cast<const float*>(scale.data_ptr()));
void concat_and_cache_mla(
torch::Tensor& kv_c, // [num_tokens, kv_lora_rank]
torch::Tensor& k_pe, // [num_tokens, pe_dim]
@ -531,20 +719,43 @@ void concat_and_cache_mla(
int pe_dim = k_pe.size(1);
int block_size = kv_cache.size(1);
TORCH_CHECK(kv_cache.size(2) == kv_lora_rank + pe_dim);
if (kv_cache_dtype == "fp8_ds_mla") {
TORCH_CHECK(kv_lora_rank == 512, "kv_lora_rank must be 512 for fp8_ds_mla");
TORCH_CHECK(pe_dim == 64, "pe_dim must be 64 for fp8_ds_mla");
TORCH_CHECK(kv_cache.size(2) == 656 / kv_cache.itemsize(),
"kv_cache.size(2) must be 656 bytes for fp8_ds_mla");
TORCH_CHECK(kv_c.itemsize() == 2,
"kv_c.itemsize() must be 2 for fp8_ds_mla");
TORCH_CHECK(k_pe.itemsize() == 2,
"k_pe.itemsize() must be 2 for fp8_ds_mla");
} else {
TORCH_CHECK(kv_cache.size(2) == kv_lora_rank + pe_dim);
}
int kv_c_stride = kv_c.stride(0);
int k_pe_stride = k_pe.stride(0);
int block_stride = kv_cache.stride(0);
int entry_stride = kv_cache.stride(1);
dim3 grid(num_tokens);
dim3 block(std::min(kv_lora_rank, 512));
const at::cuda::OptionalCUDAGuard device_guard(device_of(kv_c));
const cudaStream_t stream = at::cuda::getCurrentCUDAStream();
DISPATCH_BY_KV_CACHE_DTYPE(kv_c.dtype(), kv_cache_dtype,
CALL_CONCAT_AND_CACHE_MLA);
if (kv_cache_dtype == "fp8_ds_mla") {
dim3 grid(num_tokens);
// For the NoPE part, each tile of 128 elements is handled by half of one
// warp (16 threads). There are 4 total tiles, so 2 warps (64 threads).
// Lanes 0 and 16 of each warp write the scale values for that warp's tiles.
// The RoPE part (last 64 elements) is handled by another 1 warp (32
// threads). So in total, we use 3 warps (96 threads) per block.
dim3 block(96);
DISPATCH_BY_KV_CACHE_DTYPE(kv_c.dtype(), kv_cache_dtype,
CALL_CONCAT_AND_CACHE_DS_MLA);
} else {
dim3 grid(num_tokens);
dim3 block(std::min(kv_lora_rank, 512));
DISPATCH_BY_KV_CACHE_DTYPE(kv_c.dtype(), kv_cache_dtype,
CALL_CONCAT_AND_CACHE_MLA);
}
}
namespace vllm {
@ -922,3 +1133,42 @@ void cp_gather_cache(
TORCH_CHECK(false, "Unsupported data type width: ", dtype_bits);
}
}
// Macro to dispatch the kernel based on the data type.
#define CALL_INDEXER_K_QUANT_AND_CACHE(KV_T, CACHE_T, KV_DTYPE) \
vllm::indexer_k_quant_and_cache_kernel<KV_T, CACHE_T, KV_DTYPE> \
<<<grid, block, 0, stream>>>( \
reinterpret_cast<KV_T*>(k.data_ptr()), \
reinterpret_cast<CACHE_T*>(kv_cache.data_ptr()), \
slot_mapping.data_ptr<int64_t>(), head_dim, quant_block_size, \
cache_block_size, cache_stride, use_ue8m0);
void indexer_k_quant_and_cache(
torch::Tensor& k, // [num_tokens, head_dim]
torch::Tensor& kv_cache, // [num_blocks, block_size, cache_stride]
torch::Tensor& slot_mapping, // [num_tokens]
int64_t quant_block_size, // quantization block size
const std::string& scale_fmt) {
int num_tokens = k.size(0);
int head_dim = k.size(1);
int cache_block_size = kv_cache.size(1);
int cache_stride = kv_cache.size(2);
bool use_ue8m0 = scale_fmt == "ue8m0";
TORCH_CHECK(k.device() == kv_cache.device(),
"k and kv_cache must be on the same device");
TORCH_CHECK(k.device() == slot_mapping.device(),
"k and slot_mapping must be on the same device");
TORCH_CHECK(head_dim % quant_block_size == 0,
"head_dim must be divisible by quant_block_size");
constexpr int vec_size = 4;
dim3 grid(num_tokens, (head_dim + quant_block_size * vec_size - 1) /
(quant_block_size * vec_size));
dim3 block(32, vec_size);
const at::cuda::OptionalCUDAGuard device_guard(device_of(k));
const cudaStream_t stream = at::cuda::getCurrentCUDAStream();
DISPATCH_BY_KV_CACHE_DTYPE(k.dtype(), "fp8_e4m3",
CALL_INDEXER_K_QUANT_AND_CACHE);
}

View File

@ -0,0 +1,16 @@
#pragma once
#include <cstdlib>
#include <string>
#include <cctype>
namespace vllm {
// vllm_kernel_override_batch_invariant(); returns true
// if env VLLM_KERNEL_OVERRIDE_BATCH_INVARIANT=1
inline bool vllm_kernel_override_batch_invariant() {
std::string env_key = "VLLM_KERNEL_OVERRIDE_BATCH_INVARIANT";
const char* val = std::getenv(env_key.c_str());
return (val && std::atoi(val) != 0) ? 1 : 0;
}
} // namespace vllm

View File

@ -14,7 +14,8 @@
// arm implementation
#include "cpu_types_arm.hpp"
#else
#warning "unsupported vLLM cpu implementation"
#warning "unsupported vLLM cpu implementation, vLLM will compile with scalar"
#include "cpu_types_scalar.hpp"
#endif
#ifdef _OPENMP

View File

@ -0,0 +1,513 @@
#include <cmath>
#include <cstdint>
#include <cstring>
#include <torch/all.h>
#include "float_convert.hpp"
namespace vec_op {
#define VLLM_DISPATCH_CASE_FLOATING_TYPES(...) \
AT_DISPATCH_CASE(at::ScalarType::Float, __VA_ARGS__) \
AT_DISPATCH_CASE(at::ScalarType::BFloat16, __VA_ARGS__) \
AT_DISPATCH_CASE(at::ScalarType::Half, __VA_ARGS__)
#define VLLM_DISPATCH_FLOATING_TYPES(TYPE, NAME, ...) \
AT_DISPATCH_SWITCH(TYPE, NAME, VLLM_DISPATCH_CASE_FLOATING_TYPES(__VA_ARGS__))
#ifndef CPU_OP_GUARD
#define CPU_KERNEL_GUARD_IN(NAME)
#define CPU_KERNEL_GUARD_OUT(NAME)
#else
#define CPU_KERNEL_GUARD_IN(NAME) \
std::cout << #NAME << " invoked." << std::endl;
#define CPU_KERNEL_GUARD_OUT(NAME) \
std::cout << #NAME << " exit." << std::endl;
#endif
#define FORCE_INLINE __attribute__((always_inline)) inline
#define __max(a, b) ((a) > (b) ? (a) : (b))
#define __min(a, b) ((a) < (b) ? (a) : (b))
#define __abs(a) ((a) < (0) ? (0 - a) : (a))
typedef struct f16x8_t {
uint16_t val[8];
} f16x8_t;
typedef struct f16x16_t {
uint16_t val[16];
} f16x16_t;
typedef struct f16x32_t {
uint16_t val[32];
} f16x32_t;
typedef struct f32x4_t {
float val[4];
} f32x4_t;
typedef struct f32x8_t {
float val[8];
} f32x8_t;
typedef struct f32x16_t {
float val[16];
} f32x16_t;
namespace {
template <typename T, T... indexes, typename F>
constexpr void unroll_loop_item(std::integer_sequence<T, indexes...>, F&& f) {
(f(std::integral_constant<T, indexes>{}), ...);
};
}; // namespace
template <typename T, T count, typename F,
typename = std::enable_if_t<std::is_invocable_v<F, T> > >
constexpr void unroll_loop(F&& f) {
unroll_loop_item(std::make_integer_sequence<T, count>{}, std::forward<F>(f));
}
template <typename T>
struct Vec {
constexpr static int get_elem_num() { return T::VEC_ELEM_NUM; }
};
struct FP32Vec8;
struct FP32Vec16;
struct FP16Vec8 : public Vec<FP16Vec8> {
constexpr static int VEC_ELEM_NUM = 8;
f16x8_t reg;
explicit FP16Vec8(const void* ptr)
: reg(*reinterpret_cast<const f16x8_t*>(ptr)) {};
explicit FP16Vec8(const FP32Vec8&);
void save(void* ptr) const { *reinterpret_cast<f16x8_t*>(ptr) = reg; }
};
struct FP16Vec16 : public Vec<FP16Vec16> {
constexpr static int VEC_ELEM_NUM = 16;
f16x16_t reg;
explicit FP16Vec16(const void* ptr)
: reg(*reinterpret_cast<const f16x16_t*>(ptr)) {};
explicit FP16Vec16(const FP32Vec16&);
void save(void* ptr) const { *reinterpret_cast<f16x16_t*>(ptr) = reg; }
void save(void* ptr, const int elem_num) const {
int num = __min(elem_num, VEC_ELEM_NUM);
std::memcpy(ptr, &(reg.val[0]), num * sizeof(uint16_t));
}
};
struct BF16Vec8 : public Vec<BF16Vec8> {
constexpr static int VEC_ELEM_NUM = 8;
f16x8_t reg;
explicit BF16Vec8(const void* ptr)
: reg(*reinterpret_cast<const f16x8_t*>(ptr)) {};
explicit BF16Vec8(const FP32Vec8&);
void save(void* ptr) const { *reinterpret_cast<f16x8_t*>(ptr) = reg; }
};
struct BF16Vec16 : public Vec<BF16Vec16> {
constexpr static int VEC_ELEM_NUM = 16;
f16x16_t reg;
explicit BF16Vec16(const void* ptr)
: reg(*reinterpret_cast<const f16x16_t*>(ptr)) {};
explicit BF16Vec16(const FP32Vec16&);
void save(void* ptr) const { *reinterpret_cast<f16x16_t*>(ptr) = reg; }
void save(void* ptr, const int elem_num) const {
int num = __min(elem_num, VEC_ELEM_NUM);
std::memcpy(ptr, &(reg.val[0]), num * sizeof(uint16_t));
}
};
struct BF16Vec32 : public Vec<BF16Vec32> {
constexpr static int VEC_ELEM_NUM = 32;
f16x32_t reg;
explicit BF16Vec32(const void* ptr)
: reg(*reinterpret_cast<const f16x32_t*>(ptr)) {};
explicit BF16Vec32(f16x32_t data) : reg(data) {};
explicit BF16Vec32(BF16Vec8& vec8_data) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = vec8_data.reg.val[i % BF16Vec8::VEC_ELEM_NUM];
}
}
void save(void* ptr) const { *reinterpret_cast<f16x32_t*>(ptr) = reg; }
};
struct FP32Vec4 : public Vec<FP32Vec4> {
constexpr static int VEC_ELEM_NUM = 4;
f32x4_t reg;
explicit FP32Vec4(float v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = v;
}
}
explicit FP32Vec4() {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = 0.0f;
}
}
explicit FP32Vec4(const float* ptr)
: reg(*reinterpret_cast<const f32x4_t*>(ptr)) {};
explicit FP32Vec4(f32x4_t data) : reg(data) {};
explicit FP32Vec4(const FP32Vec4& data) : reg(data.reg) {};
};
struct FP32Vec8 : public Vec<FP32Vec8> {
constexpr static int VEC_ELEM_NUM = 8;
f32x8_t reg;
explicit FP32Vec8(float v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = v;
}
}
explicit FP32Vec8() {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = 0.0f;
}
}
explicit FP32Vec8(const float* ptr)
: reg(*reinterpret_cast<const f32x8_t*>(ptr)) {};
explicit FP32Vec8(f32x8_t data) : reg(data) {};
explicit FP32Vec8(const FP32Vec8& data) : reg(data.reg) {};
explicit FP32Vec8(const FP16Vec8& v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = fp16_to_float(v.reg.val[i]);
}
}
FP32Vec8(const BF16Vec8& v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = bf16_to_float(v.reg.val[i]);
}
}
float reduce_sum() const {
float result = 0;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result += reg.val[i];
}
return result;
}
FP32Vec8 exp() const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = expf(reg.val[i]);
}
return FP32Vec8(ret);
}
FP32Vec8 tanh() const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = tanhf(reg.val[i]);
}
return FP32Vec8(ret);
}
FP32Vec8 er() const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = erf(reg.val[i]);
}
return FP32Vec8(ret);
}
FP32Vec8 operator*(const FP32Vec8& b) const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = reg.val[i] * b.reg.val[i];
}
return FP32Vec8(ret);
}
FP32Vec8 operator+(const FP32Vec8& b) const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = reg.val[i] + b.reg.val[i];
}
return FP32Vec8(ret);
}
FP32Vec8 operator-(const FP32Vec8& b) const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = reg.val[i] - b.reg.val[i];
}
return FP32Vec8(ret);
}
FP32Vec8 operator/(const FP32Vec8& b) const {
f32x8_t ret;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
ret.val[i] = reg.val[i] / b.reg.val[i];
}
return FP32Vec8(ret);
}
void save(void* ptr) const { *reinterpret_cast<f32x8_t*>(ptr) = reg; }
};
struct FP32Vec16 : public Vec<FP32Vec16> {
constexpr static int VEC_ELEM_NUM = 16;
f32x16_t reg;
explicit FP32Vec16(float v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = v;
}
}
explicit FP32Vec16() {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = 0.0f;
}
}
explicit FP32Vec16(const float* ptr)
: reg(*reinterpret_cast<const f32x16_t*>(ptr)) {};
explicit FP32Vec16(f32x16_t data) : reg(data) {};
FP32Vec16(const FP32Vec4& data) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = data.reg.val[i % FP32Vec4::VEC_ELEM_NUM];
}
}
FP32Vec16(const FP32Vec8& data) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = data.reg.val[i % FP32Vec8::VEC_ELEM_NUM];
}
}
FP32Vec16(const FP32Vec16& data) : reg(data.reg) {};
explicit FP32Vec16(const FP16Vec16& v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = fp16_to_float(v.reg.val[i]);
}
}
explicit FP32Vec16(const BF16Vec16& v) {
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
reg.val[i] = bf16_to_float(v.reg.val[i]);
}
}
explicit FP32Vec16(const FP16Vec8& v) : FP32Vec16(FP32Vec8(v)) {};
FP32Vec16(const BF16Vec8& v) : FP32Vec16(FP32Vec8(v)) {};
FP32Vec16 operator*(const FP32Vec16& b) const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = reg.val[i] * b.reg.val[i];
}
return result;
}
FP32Vec16 operator+(const FP32Vec16& b) const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = reg.val[i] + b.reg.val[i];
}
return result;
}
FP32Vec16 operator-(const FP32Vec16& b) const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = reg.val[i] - b.reg.val[i];
}
return result;
}
FP32Vec16 operator/(const FP32Vec16& b) const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = reg.val[i] / b.reg.val[i];
}
return result;
}
FP32Vec16 max(const FP32Vec16& b) const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = __max(reg.val[i], b.reg.val[i]);
}
return result;
}
FP32Vec16 min(const FP32Vec16& b) const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = __min(reg.val[i], b.reg.val[i]);
}
return result;
}
FP32Vec16 abs() const {
FP32Vec16 result(0.0f);
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result.reg.val[i] = __abs(reg.val[i]);
}
return result;
}
float reduce_sum() const {
float result = 0.0f;
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result += reg.val[i];
}
return result;
}
float reduce_max() const {
float result = reg.val[0];
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result = __max(reg.val[i], result);
}
return result;
}
float reduce_min() const {
float result = reg.val[0];
for (int i = 0; i < VEC_ELEM_NUM; ++i) {
result = __min(reg.val[i], result);
}
return result;
}
template <int group_size>
float reduce_sub_sum(int idx) {
static_assert(VEC_ELEM_NUM % group_size == 0);
float sum = 0.0;
int start = idx * group_size;
int end = (idx + 1) * group_size;
for (; (start < VEC_ELEM_NUM) && (start < end); ++start) {
sum += reg.val[start];
}
return sum;
}
void save(void* ptr) const { *reinterpret_cast<f32x16_t*>(ptr) = reg; }
};
template <typename T>
struct VecType {
using vec_type = void;
};
template <typename T>
using vec_t = typename VecType<T>::vec_type;
template <>
struct VecType<float> {
using vec_type = FP32Vec8;
};
template <>
struct VecType<c10::Half> {
using vec_type = FP16Vec8;
};
template <>
struct VecType<c10::BFloat16> {
using vec_type = BF16Vec8;
};
template <typename T>
void storeFP32(float v, T* ptr) {
*ptr = v;
}
/*
template <> inline void storeFP32<c10::Half>(float v, c10::Half *ptr) {
c10::Half __attribute__((__may_alias__)) *v_ptr =
reinterpret_cast<c10::Half *>(&v);
*ptr = *(v_ptr + 1);
}
*/
template <>
inline void storeFP32<c10::Half>(float v, c10::Half* ptr) {
uint16_t fp16 = float_to_fp16(v);
*reinterpret_cast<uint16_t*>(ptr) = fp16;
}
template <>
inline void storeFP32<c10::BFloat16>(float v, c10::BFloat16* ptr) {
c10::BFloat16 __attribute__((__may_alias__))* v_ptr =
reinterpret_cast<c10::BFloat16*>(&v);
*ptr = *(v_ptr + 1);
}
inline FP16Vec16::FP16Vec16(const FP32Vec16& v) {
int i = 0;
for (i = 0; i < FP16Vec16::VEC_ELEM_NUM; ++i) {
reg.val[i] = float_to_fp16(v.reg.val[i]);
}
}
inline FP16Vec8 ::FP16Vec8(const FP32Vec8& v) {
int i = 0;
for (i = 0; i < FP16Vec8::VEC_ELEM_NUM; ++i) {
reg.val[i] = float_to_fp16(v.reg.val[i]);
}
}
inline void fma(FP32Vec16& acc, FP32Vec16& a, FP32Vec16& b) {
acc = acc + a * b;
}
inline BF16Vec8::BF16Vec8(const FP32Vec8& v) {
int i = 0;
for (i = 0; i < BF16Vec8::VEC_ELEM_NUM; ++i) {
reg.val[i] = float_to_bf16(v.reg.val[i]);
}
}
inline BF16Vec16::BF16Vec16(const FP32Vec16& v) {
int i = 0;
for (i = 0; i < BF16Vec16::VEC_ELEM_NUM; ++i) {
reg.val[i] = float_to_bf16(v.reg.val[i]);
}
}
inline void prefetch(const void* addr) { __builtin_prefetch(addr, 0, 3); }
}; // namespace vec_op

106
csrc/cpu/float_convert.hpp Normal file
View File

@ -0,0 +1,106 @@
static float bf16_to_float(uint16_t bf16) {
uint32_t bits = static_cast<uint32_t>(bf16) << 16;
float fp32;
std::memcpy(&fp32, &bits, sizeof(fp32));
return fp32;
}
static uint16_t float_to_bf16(float fp32) {
uint32_t bits;
std::memcpy(&bits, &fp32, sizeof(fp32));
return static_cast<uint16_t>(bits >> 16);
}
/************************************************
* Copyright (c) 2015 Princeton Vision Group
* Licensed under the MIT license.
* Codes below copied from
* https://github.com/PrincetonVision/marvin/tree/master/tools/tensorIO_matlab
*************************************************/
static uint16_t float_to_fp16(float fp32) {
uint16_t fp16;
unsigned x;
unsigned u, remainder, shift, lsb, lsb_s1, lsb_m1;
unsigned sign, exponent, mantissa;
std::memcpy(&x, &fp32, sizeof(fp32));
u = (x & 0x7fffffff);
// Get rid of +NaN/-NaN case first.
if (u > 0x7f800000) {
fp16 = 0x7fffU;
return fp16;
}
sign = ((x >> 16) & 0x8000);
// Get rid of +Inf/-Inf, +0/-0.
if (u > 0x477fefff) {
fp16 = sign | 0x7c00U;
return fp16;
}
if (u < 0x33000001) {
fp16 = (sign | 0x0000);
return fp16;
}
exponent = ((u >> 23) & 0xff);
mantissa = (u & 0x7fffff);
if (exponent > 0x70) {
shift = 13;
exponent -= 0x70;
} else {
shift = 0x7e - exponent;
exponent = 0;
mantissa |= 0x800000;
}
lsb = (1 << shift);
lsb_s1 = (lsb >> 1);
lsb_m1 = (lsb - 1);
// Round to nearest even.
remainder = (mantissa & lsb_m1);
mantissa >>= shift;
if (remainder > lsb_s1 || (remainder == lsb_s1 && (mantissa & 0x1))) {
++mantissa;
if (!(mantissa & 0x3ff)) {
++exponent;
mantissa = 0;
}
}
fp16 = (sign | (exponent << 10) | mantissa);
return fp16;
}
static float fp16_to_float(uint16_t fp16) {
unsigned sign = ((fp16 >> 15) & 1);
unsigned exponent = ((fp16 >> 10) & 0x1f);
unsigned mantissa = ((fp16 & 0x3ff) << 13);
int temp;
float fp32;
if (exponent == 0x1f) { /* NaN or Inf */
mantissa = (mantissa ? (sign = 0, 0x7fffff) : 0);
exponent = 0xff;
} else if (!exponent) { /* Denorm or Zero */
if (mantissa) {
unsigned int msb;
exponent = 0x71;
do {
msb = (mantissa & 0x400000);
mantissa <<= 1; /* normalize */
--exponent;
} while (!msb);
mantissa &= 0x7fffff; /* 1.mantissa is implicit */
}
} else {
exponent += 0x70;
}
temp = ((sign << 31) | (exponent << 23) | mantissa);
std::memcpy(&fp32, &temp, sizeof(temp));
return fp32;
}

View File

@ -88,8 +88,18 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
" int tp_rank, int blocksparse_local_blocks,"
" int blocksparse_vert_stride, int blocksparse_block_size,"
" int blocksparse_head_sliding_step) -> ()");
ops.impl("paged_attention_v1", torch::kCPU, &paged_attention_v1);
ops.def(
"dynamic_4bit_int_moe("
"Tensor x, Tensor topk_ids, Tensor topk_weights,"
"Tensor w13_packed, Tensor w2_packed, int H, int I, int I2,"
"int group_size, bool apply_router_weight_on_input, int activation_kind"
") -> Tensor");
ops.impl("dynamic_4bit_int_moe", torch::kCPU, &dynamic_4bit_int_moe_cpu);
// PagedAttention V2.
ops.def(
"paged_attention_v2("

View File

@ -0,0 +1,64 @@
#pragma once
#include <cuda_runtime_api.h>
#include <algorithm>
// maximum blocks per SM cap
#ifndef VLLM_LAUNCH_BLOCKS_CAP
#define VLLM_LAUNCH_BLOCKS_CAP 4
#endif
// Compile-time estimate of max threads per SM for launch bounds.
// Families: 1024, 1536, 2048 threads/SM.
#ifndef VLLM_MAX_THREADS_PER_SM
#ifdef __CUDA_ARCH__
/* 1024 thr/SM: Turing (sm_75) */
#if (__CUDA_ARCH__ == 750)
#define VLLM_MAX_THREADS_PER_SM 1024
/* 1536 thr/SM: Ampere GA10x (sm_86/87), Ada (sm_89),
GB20x consumer (sm_120/121), Thor (sm_101 or sm_110) */
#elif (__CUDA_ARCH__ == 860) || (__CUDA_ARCH__ == 870) || \
(__CUDA_ARCH__ == 890) || (__CUDA_ARCH__ == 1010) || \
(__CUDA_ARCH__ == 1100) || (__CUDA_ARCH__ == 1200) || \
(__CUDA_ARCH__ == 1210)
#define VLLM_MAX_THREADS_PER_SM 1536
/* 2048 thr/SM: Volta (sm_70/72), Ampere GA100 (sm_80),
Hopper (sm_90), Blackwell (sm_100/103) */
#elif (__CUDA_ARCH__ == 700) || (__CUDA_ARCH__ == 720) || \
(__CUDA_ARCH__ == 800) || (__CUDA_ARCH__ == 900) || \
(__CUDA_ARCH__ == 1000) || (__CUDA_ARCH__ == 1030)
#define VLLM_MAX_THREADS_PER_SM 2048
/* Fallback: use 2048 for unknown future CCs */
#else
#define VLLM_MAX_THREADS_PER_SM 2048
#endif
#else
/* Host pass (no __CUDA_ARCH__): neutral default */
#define VLLM_MAX_THREADS_PER_SM 2048
#endif
#endif
// compute the number of blocks per SM to request in __launch_bounds__
#define VLLM_BLOCKS_DIV(VAL) (VLLM_MAX_THREADS_PER_SM / (VAL))
#define VLLM_CLAMP_BLOCKS_PER_SM(VAL) \
(((VAL) <= 0) \
? 1 \
: (((VAL) < VLLM_LAUNCH_BLOCKS_CAP) ? (VAL) : VLLM_LAUNCH_BLOCKS_CAP))
#define VLLM_BLOCKS_PER_SM(BLOCK_THREADS) \
VLLM_CLAMP_BLOCKS_PER_SM(VLLM_BLOCKS_DIV(BLOCK_THREADS))
// runtime-time helper to compute blocks/SM
static inline int vllm_runtime_blocks_per_sm(int block_threads) {
int device = -1;
cudaGetDevice(&device);
int max_threads_per_sm = VLLM_MAX_THREADS_PER_SM;
cudaDeviceGetAttribute(&max_threads_per_sm,
cudaDevAttrMaxThreadsPerMultiProcessor, device);
int blocks = (block_threads > 0) ? (max_threads_per_sm / block_threads) : 1;
return VLLM_CLAMP_BLOCKS_PER_SM(blocks);
}

View File

@ -1,6 +1,7 @@
#include "type_convert.cuh"
#include "dispatch_utils.h"
#include "cub_helpers.h"
#include "core/batch_invariant.hpp"
#include <torch/cuda.h>
#include <c10/cuda/CUDAGuard.h>
@ -413,7 +414,9 @@ void fused_add_rms_norm(torch::Tensor& input, // [..., hidden_size]
wt_ptr % req_alignment_bytes == 0;
bool offsets_are_multiple_of_vector_width =
hidden_size % vector_width == 0 && input_stride % vector_width == 0;
if (ptrs_are_aligned && offsets_are_multiple_of_vector_width) {
bool batch_invariant_launch = vllm::vllm_kernel_override_batch_invariant();
if (ptrs_are_aligned && offsets_are_multiple_of_vector_width &&
!batch_invariant_launch) {
LAUNCH_FUSED_ADD_RMS_NORM(8);
} else {
LAUNCH_FUSED_ADD_RMS_NORM(0);
@ -459,7 +462,8 @@ void poly_norm(torch::Tensor& out, // [..., hidden_size]
auto inp_ptr = reinterpret_cast<std::uintptr_t>(input.data_ptr());
auto out_ptr = reinterpret_cast<std::uintptr_t>(out.data_ptr());
bool ptrs_are_aligned = inp_ptr % 16 == 0 && out_ptr % 16 == 0;
if (ptrs_are_aligned && hidden_size % 8 == 0) {
bool batch_invariant_launch = vllm::vllm_kernel_override_batch_invariant();
if (ptrs_are_aligned && hidden_size % 8 == 0 && !batch_invariant_launch) {
LAUNCH_FUSED_POLY_NORM(8);
} else {
LAUNCH_FUSED_POLY_NORM(0);

View File

@ -9,6 +9,7 @@
#include "quantization/fp8/common.cuh"
#include "dispatch_utils.h"
#include "cub_helpers.h"
#include "core/batch_invariant.hpp"
#include <torch/cuda.h>
#include <c10/cuda/CUDAGuard.h>
@ -240,7 +241,9 @@ void fused_add_rms_norm_static_fp8_quant(
auto wt_ptr = reinterpret_cast<std::uintptr_t>(weight.data_ptr());
bool ptrs_are_aligned =
inp_ptr % 16 == 0 && res_ptr % 16 == 0 && wt_ptr % 16 == 0;
if (ptrs_are_aligned && hidden_size % 8 == 0 && input_stride % 8 == 0) {
bool batch_invariant_launch = vllm::vllm_kernel_override_batch_invariant();
if (ptrs_are_aligned && hidden_size % 8 == 0 && input_stride % 8 == 0 &&
!batch_invariant_launch) {
LAUNCH_FUSED_ADD_RMS_NORM(8);
} else {
LAUNCH_FUSED_ADD_RMS_NORM(0);

View File

@ -0,0 +1,156 @@
#include <ATen/ATen.h>
#include <ATen/Parallel.h>
#include <torch/all.h>
// _dyn_quant_matmul_4bit is only available on AArch64.
#if defined(__aarch64__)
#include <ATen/ops/_dyn_quant_matmul_4bit.h>
#endif
inline torch::Tensor mm(const torch::Tensor& a, const torch::Tensor& packed_w,
int64_t group_size_eff, int64_t in_features,
int64_t out_features) {
#if defined(__aarch64__)
return at::_ops::_dyn_quant_matmul_4bit::call(a, packed_w, group_size_eff,
in_features, out_features);
#else
TORCH_CHECK(false,
"dynamic 4-bit int MoE path requires AArch64 (ARM64); "
"_dyn_quant_matmul_4bit is unavailable on this architecture");
return {};
#endif
}
enum ActivationKind : int64_t {
SwiGLU_Gu = 0, // act = SiLU(g) * u
SwiGLUOAI = 1, // act = SiLU(u) * g
SiLU = 2 // SiLU
};
torch::Tensor dynamic_4bit_int_moe_cpu(
torch::Tensor x, torch::Tensor topk_ids, torch::Tensor topk_weights,
torch::Tensor w13_packed, torch::Tensor w2_packed, int64_t H, int64_t I,
int64_t I2, int64_t group_size, bool apply_router_weight_on_input,
int64_t activation_kind) {
TORCH_CHECK(x.dim() == 2, "x must be 2D");
TORCH_CHECK(topk_ids.dim() == 2 && topk_weights.dim() == 2,
"topk tensors must be [T, K]");
TORCH_CHECK(
w13_packed.size(0) == w2_packed.size(0),
"w13_packed and w2_packed must have same number of experts in dim 0");
TORCH_CHECK(I2 == 2 * I, "I2 must equal 2*I");
const int64_t T = x.size(0);
const int64_t K = topk_ids.size(1);
const int64_t E = w13_packed.size(0);
const int64_t N = T * K;
auto x_c = x.contiguous();
auto ids_c = topk_ids.contiguous();
auto gates_c = topk_weights.to(at::kFloat).contiguous();
// bucketing tokens -> experts
c10::SmallVector<int64_t, 64> counts(
E, 0); // Small vector uses stack allocation
{
const auto* ids_ptr = ids_c.data_ptr<int64_t>();
for (int64_t i = 0; i < N; ++i) {
const int64_t e_id = ids_ptr[i];
TORCH_CHECK(0 <= e_id && e_id < E, "expert id out of range");
counts[e_id]++;
}
}
c10::SmallVector<int64_t, 65> offsets(E + 1, 0); // ( E +1 )
for (int64_t e = 0; e < E; ++e) offsets[e + 1] = offsets[e] + counts[e];
auto expert_tokens = at::empty({offsets[E]}, ids_c.options());
auto expert_gates = at::empty({offsets[E]}, gates_c.options());
{
c10::SmallVector<int64_t, 64> cursor(E, 0);
const auto* ids_ptr = ids_c.data_ptr<int64_t>();
const auto* gts_ptr = gates_c.data_ptr<float>();
auto* tok_ptr = expert_tokens.data_ptr<int64_t>();
auto* gate_ptr = expert_gates.data_ptr<float>();
for (int64_t t = 0; t < T; ++t) {
const int64_t base = t * K;
for (int64_t k = 0; k < K; ++k) {
const int64_t idx = base + k;
const int64_t e = ids_ptr[idx];
const int64_t p = offsets[e] + (cursor[e]++);
tok_ptr[p] = t;
gate_ptr[p] = gts_ptr[idx];
}
}
}
const int64_t g_eff_13 = (group_size != -1) ? group_size : H;
const int64_t g_eff_2 = (group_size != -1) ? group_size : I;
// Per-expert outputs filled in parallel
std::vector<torch::Tensor> y_list(E);
y_list.resize(E);
at::parallel_for(0, E, 1, [&](int64_t e_begin, int64_t e_end) {
for (int64_t e = e_begin; e < e_end; ++e) {
const int64_t te = counts[e];
if (te == 0) {
y_list[e] = at::empty({0, H}, x_c.options());
continue;
}
const int64_t start = offsets[e];
auto sel_tokens =
expert_tokens.narrow(/*dim=*/0, /*start=*/start, /*length=*/te);
auto gates_e =
expert_gates.narrow(/*dim=*/0, /*start=*/start, /*length=*/te);
auto x_e = x_c.index_select(/*dim=*/0, sel_tokens);
if (apply_router_weight_on_input) {
x_e = x_e.mul(gates_e.unsqueeze(1));
}
auto w13_e = w13_packed.select(/*dim=*/0, e);
auto w2_e = w2_packed.select(/*dim=*/0, e);
// W13
auto y13 =
mm(x_e, w13_e, g_eff_13, /*in_features=*/H, /*out_features=*/I2);
auto g_part = y13.narrow(/*dim=*/1, /*start=*/0, /*length=*/I);
auto u_part = y13.narrow(/*dim=*/1, /*start=*/I, /*length=*/I);
torch::Tensor act;
if (activation_kind == ActivationKind::SwiGLUOAI) { // SwiGLUOAI
constexpr double kAlpha = 1.702; // GPT-OSS default
constexpr double kLimit = 7.0; // GPT-OSS default
auto gate_c = at::clamp_max(g_part, kLimit);
auto up_c = at::clamp(u_part, -kLimit, kLimit);
auto glu = gate_c.mul(at::sigmoid(gate_c.mul(kAlpha)));
act = up_c.add(1.0).mul(glu);
} else { // SiLU , SwiGLU_GU, vLLM maps silu to SiluAndMul()
act = at::silu(g_part).mul(u_part);
}
// W2
auto y = mm(act, w2_e, g_eff_2, /*in_features=*/I, /*out_features=*/H);
if (!apply_router_weight_on_input) {
y = y.mul(gates_e.unsqueeze(1));
}
// Store per-expert result
y_list[e] = y;
}
});
// Concatenate all expert outputs to match expert_tokens order
auto Y_all = at::cat(y_list, /*dim=*/0);
auto out = at::zeros({T, H}, x.options());
out =
at::index_add(out, /*dim=*/0, /*index=*/expert_tokens, /*source=*/Y_all);
return out;
}

View File

@ -418,6 +418,15 @@ __device__ inline T neg_inf() {
return cuda_cast<T, float>(-cuda::std::numeric_limits<float>::infinity());
}
template <typename T>
__device__ inline bool is_finite(const T val) {
#if (__CUDACC_VER_MAJOR__ * 10000 + __CUDACC_VER_MINOR__ * 100 >= 120800)
return cuda::std::isfinite(val);
#else
return isfinite(cuda_cast<float, T>(val));
#endif
}
template <typename T>
__device__ void topk_with_k2(T* output, T const* input,
cg::thread_block_tile<32> const& tile,
@ -533,7 +542,7 @@ __global__ void group_idx_and_topk_idx_kernel(
// calculate group_idx
int32_t target_num_min = WARP_SIZE - n_group + topk_group;
// The check is necessary to avoid abnormal input
if (lane_id < n_group && cuda::std::isfinite(group_scores[lane_id])) {
if (lane_id < n_group && is_finite(group_scores[lane_id])) {
value = group_scores[lane_id];
}
@ -568,11 +577,10 @@ __global__ void group_idx_and_topk_idx_kernel(
int32_t offset = i_group * num_experts_per_group;
for (int32_t i = lane_id; i < align_num_experts_per_group;
i += WARP_SIZE) {
T candidates =
(i < num_experts_per_group) &&
cuda::std::isfinite(scores_with_bias[offset + i])
? scores_with_bias[offset + i]
: neg_inf<T>();
T candidates = (i < num_experts_per_group) &&
is_finite(scores_with_bias[offset + i])
? scores_with_bias[offset + i]
: neg_inf<T>();
queue.add(candidates, offset + i);
}
if (group_scores[i_group] == topk_group_value) {

View File

@ -44,6 +44,9 @@ __global__ void moe_align_block_size_kernel(
for (size_t i = tid; i < numel; i += stride) {
int expert_id = topk_ids[i];
if (expert_id >= num_experts) {
continue;
}
int warp_idx = expert_id / experts_per_warp;
int expert_offset = expert_id % experts_per_warp;
atomicAdd(&shared_counts[warp_idx * experts_per_warp + expert_offset], 1);
@ -95,12 +98,15 @@ template <typename scalar_t>
__global__ void count_and_sort_expert_tokens_kernel(
const scalar_t* __restrict__ topk_ids,
int32_t* __restrict__ sorted_token_ids, int32_t* __restrict__ cumsum_buffer,
size_t numel) {
size_t numel, int32_t num_experts) {
const size_t tid = blockIdx.x * blockDim.x + threadIdx.x;
const size_t stride = blockDim.x * gridDim.x;
for (size_t i = tid; i < numel; i += stride) {
int32_t expert_id = topk_ids[i];
if (expert_id >= num_experts) {
continue;
}
int32_t rank_post_pad = atomicAdd(&cumsum_buffer[expert_id], 1);
sorted_token_ids[rank_post_pad] = i;
}
@ -269,7 +275,7 @@ void moe_align_block_size(torch::Tensor topk_ids, int64_t num_experts,
sort_kernel<<<actual_blocks, block_threads, 0, stream>>>(
topk_ids.data_ptr<scalar_t>(),
sorted_token_ids.data_ptr<int32_t>(),
cumsum_buffer.data_ptr<int32_t>(), topk_ids.numel());
cumsum_buffer.data_ptr<int32_t>(), topk_ids.numel(), num_experts);
}
});
}

View File

@ -21,6 +21,7 @@
#include <c10/cuda/CUDAGuard.h>
#include "../cuda_compat.h"
#include "../cub_helpers.h"
#include "../core/batch_invariant.hpp"
#define MAX(a, b) ((a) > (b) ? (a) : (b))
#define MIN(a, b) ((a) < (b) ? (a) : (b))
@ -405,7 +406,8 @@ void topkGatingSoftmaxLauncherHelper(const float* input, const bool* finished, f
using Constants = detail::TopkConstants<EXPERTS, BYTES_PER_LDG, WARP_SIZE_PARAM>;
static constexpr int VPT = Constants::VPT;
static constexpr int ROWS_PER_WARP = Constants::ROWS_PER_WARP;
const int num_warps = (num_rows + ROWS_PER_WARP - 1) / ROWS_PER_WARP;
const bool batch_invariant_launch = vllm::vllm_kernel_override_batch_invariant();
const int num_warps = batch_invariant_launch ? 32 : (num_rows + ROWS_PER_WARP - 1) / ROWS_PER_WARP;
const int num_blocks = (num_warps + WARPS_PER_TB - 1) / WARPS_PER_TB;
dim3 block_dim(WARP_SIZE_PARAM, WARPS_PER_TB);

View File

@ -328,6 +328,12 @@ void selective_scan_fwd(const torch::Tensor& u, const torch::Tensor& delta,
const std::optional<torch::Tensor>& has_initial_state,
const torch::Tensor& ssm_states, int64_t pad_slot_id);
torch::Tensor dynamic_4bit_int_moe_cpu(
torch::Tensor x, torch::Tensor topk_ids, torch::Tensor topk_weights,
torch::Tensor w13_packed, torch::Tensor w2_packed, int64_t H, int64_t I,
int64_t I2, int64_t group_size, bool apply_router_weight_on_input,
int64_t activation_kind);
using fptr_t = int64_t;
fptr_t init_custom_ar(const std::vector<int64_t>& fake_ipc_ptrs,
torch::Tensor& rank_data, int64_t rank,

View File

@ -23,9 +23,14 @@
typedef __hip_bfloat162 __nv_bfloat162;
typedef __hip_bfloat16 __nv_bfloat16;
typedef __hip_bfloat16_raw __nv_bfloat16_raw;
#if defined(HIP_FP8_TYPE_OCP)
typedef __hip_fp8_e4m3 __nv_fp8_e4m3;
typedef __hip_fp8x4_e4m3 __nv_fp8x4_e4m3;
#else
// ROCm 6.2 fallback: only *_fnuz types exist
typedef __hip_fp8_e4m3_fnuz __nv_fp8_e4m3;
typedef __hip_fp8x4_e4m3_fnuz __nv_fp8x4_e4m3;
#endif
#endif
#include "core/registration.h"

View File

@ -231,7 +231,7 @@ void cutlass_gemm_blockwise_sm100_fp8_dispatch(torch::Tensor& out,
} else {
cutlass_gemm_caller_blockwise<cutlass_3x_gemm_fp8_blockwise<
OutType, 1, TILE_N, TILE_K, Shape<_64, Int<TILE_N>, Int<TILE_K>>,
Shape<_1, _1, _1>, cutlass::epilogue::NoSmemWarpSpecialized1Sm,
Shape<_1, _1, _1>, cutlass::epilogue::BlockwiseNoSmemWarpSpecialized1Sm,
cutlass::gemm::KernelTmaWarpSpecializedBlockwise1SmSm100>>(
out, a, b, a_scales, b_scales);
}
@ -245,7 +245,7 @@ void cutlass_gemm_blockwise_sm100_fp8_dispatch(torch::Tensor& out,
} else {
cutlass_gemm_caller_blockwise<cutlass_3x_gemm_fp8_blockwise<
OutType, 1, TILE_N, TILE_K, Shape<_128, Int<TILE_N>, Int<TILE_K>>,
Shape<_1, _1, _1>, cutlass::epilogue::NoSmemWarpSpecialized1Sm,
Shape<_1, _1, _1>, cutlass::epilogue::BlockwiseNoSmemWarpSpecialized1Sm,
cutlass::gemm::KernelTmaWarpSpecializedBlockwise1SmSm100>>(
out, a, b, a_scales, b_scales);
}
@ -259,7 +259,7 @@ void cutlass_gemm_blockwise_sm100_fp8_dispatch(torch::Tensor& out,
} else {
cutlass_gemm_caller_blockwise<cutlass_3x_gemm_fp8_blockwise<
OutType, 1, TILE_N, TILE_K, Shape<_256, Int<TILE_N>, Int<TILE_K>>,
Shape<_2, _1, _1>, cutlass::epilogue::NoSmemWarpSpecialized2Sm,
Shape<_2, _1, _1>, cutlass::epilogue::BlockwiseNoSmemWarpSpecialized2Sm,
cutlass::gemm::KernelTmaWarpSpecializedBlockwise2SmSm100>>(
out, a, b, a_scales, b_scales);
}
@ -271,10 +271,10 @@ void cutlass_gemm_blockwise_sm100_fp8_dispatch(torch::Tensor& out,
// TMA epilogue isn't compatible with Swap A/B
cutlass_gemm_caller_blockwise<cutlass_3x_gemm_fp8_blockwise<
OutType, TILE_M, 1, TILE_K, Shape<Int<TILE_M>, Int<TILE_N>, Int<TILE_K>>,
Shape<_1, _1, _1>, cutlass::epilogue::NoSmemWarpSpecialized1Sm,
Shape<_1, _1, _1>, cutlass::epilogue::BlockwiseNoSmemWarpSpecialized1Sm,
cutlass::gemm::KernelTmaWarpSpecializedBlockwise1SmSm100, true>>(
out, a, b, a_scales, b_scales);
}
}
} // namespace vllm
} // namespace vllm

View File

@ -25,7 +25,10 @@ void dispatch_scaled_mm(torch::Tensor& c, torch::Tensor const& a,
if constexpr (!std::is_same_v<Int8Func, std::nullptr_t>) {
int8_func(c, a, b, a_scales, b_scales, bias);
} else {
TORCH_CHECK(false, "Int8 not supported for this architecture");
int32_t version_num = get_sm_version_num();
TORCH_CHECK(
false, "Int8 not supported on SM", version_num,
". Use FP8 quantization instead, or run on older arch (SM < 100).");
}
}
} else {

View File

@ -133,4 +133,4 @@ void cutlass_scaled_mm_sm100_fp8_epilogue(torch::Tensor& out,
}
}
} // namespace vllm
} // namespace vllm

View File

@ -67,8 +67,9 @@ void cutlass_scaled_mm_sm100(torch::Tensor& c, torch::Tensor const& a,
std::optional<torch::Tensor> const& bias);
#endif
#if defined(ENABLE_SCALED_MM_SM90) && ENABLE_SCALED_MM_SM90 || \
defined(ENABLE_SCALED_MM_SM100) && ENABLE_SCALED_MM_SM100
#if defined(ENABLE_SCALED_MM_SM90) && ENABLE_SCALED_MM_SM90 || \
defined(ENABLE_SCALED_MM_SM100) && ENABLE_SCALED_MM_SM100 || \
defined(ENABLE_SCALED_MM_SM120) && ENABLE_SCALED_MM_SM120
void get_cutlass_moe_mm_data_caller(
const torch::Tensor& topk_ids, torch::Tensor& expert_offsets,
torch::Tensor& problem_sizes1, torch::Tensor& problem_sizes2,
@ -253,7 +254,7 @@ void cutlass_moe_mm(
bool per_act_token, bool per_out_ch) {
int32_t version_num = get_sm_version_num();
#if defined ENABLE_CUTLASS_MOE_SM100 && ENABLE_CUTLASS_MOE_SM100
if (version_num >= 100) {
if (version_num >= 100 && version_num < 110) {
cutlass_moe_mm_sm100(out_tensors, a_tensors, b_tensors, a_scales, b_scales,
expert_offsets, problem_sizes, a_strides, b_strides,
c_strides, per_act_token, per_out_ch);
@ -261,7 +262,7 @@ void cutlass_moe_mm(
}
#endif
#if defined ENABLE_CUTLASS_MOE_SM90 && ENABLE_CUTLASS_MOE_SM90
if (version_num >= 90) {
if (version_num >= 90 && version_num < 100) {
cutlass_moe_mm_sm90(out_tensors, a_tensors, b_tensors, a_scales, b_scales,
expert_offsets, problem_sizes, a_strides, b_strides,
c_strides, per_act_token, per_out_ch);

View File

@ -26,6 +26,7 @@
#include "dispatch_utils.h"
#include "cuda_utils.h"
#include "launch_bounds_utils.h"
#include "nvfp4_utils.cuh"
namespace vllm {
@ -63,7 +64,7 @@ __inline__ __device__ PackedVec<Type> compute_silu_mul(PackedVec<Type>& vec,
// Use UE4M3 by default.
template <class Type, bool UE8M0_SF = false>
__global__ void __launch_bounds__(1024, 4)
__global__ void __launch_bounds__(1024, VLLM_BLOCKS_PER_SM(1024))
silu_mul_cvt_fp16_to_fp4(int32_t numRows, int32_t numCols, Type const* in,
float const* SFScale, uint32_t* out,
uint32_t* SFout) {
@ -131,7 +132,8 @@ void silu_and_mul_nvfp4_quant_sm1xxa(torch::Tensor& output, // [..., d]
const at::cuda::OptionalCUDAGuard device_guard(device_of(input));
auto stream = at::cuda::getCurrentCUDAStream(input.get_device());
dim3 block(std::min(int(n / ELTS_PER_THREAD), 1024));
int const numBlocksPerSM = 2048 / block.x;
int const numBlocksPerSM =
vllm_runtime_blocks_per_sm(static_cast<int>(block.x));
dim3 grid(std::min(int(m), multiProcessorCount * numBlocksPerSM));
VLLM_DISPATCH_HALF_TYPES(

View File

@ -14,6 +14,8 @@
* limitations under the License.
*/
#include "core/registration.h"
#include <torch/all.h>
#include <cutlass/arch/arch.h>
@ -418,3 +420,7 @@ void cutlass_fp4_group_mm(
"12.8 or above.");
#endif
}
TORCH_LIBRARY_IMPL_EXPAND(TORCH_EXTENSION_NAME, CUDA, m) {
m.impl("cutlass_fp4_group_mm", &cutlass_fp4_group_mm);
}

View File

@ -26,12 +26,13 @@
#include "dispatch_utils.h"
#include "nvfp4_utils.cuh"
#include "launch_bounds_utils.h"
namespace vllm {
// Use UE4M3 by default.
template <class Type, bool UE8M0_SF = false, bool SMALL_NUM_EXPERTS = false>
__global__ void __launch_bounds__(512, 4)
__global__ void __launch_bounds__(512, VLLM_BLOCKS_PER_SM(512))
cvt_fp16_to_fp4(int32_t numRows, int32_t numCols, Type const* in,
float const* SFScale, uint32_t* out, uint32_t* SFout,
uint32_t* input_offset_by_experts,
@ -129,7 +130,7 @@ __global__ void __launch_bounds__(512, 4)
// Kernel for LARGE_M_TOPK = true (large m_topk optimized version)
template <class Type, bool UE8M0_SF = false, bool SMALL_NUM_EXPERTS = false>
__global__ void __launch_bounds__(1024, 4)
__global__ void __launch_bounds__(1024, VLLM_BLOCKS_PER_SM(1024))
cvt_fp16_to_fp4(int32_t numRows, int32_t numCols, Type const* in,
float const* SFScale, uint32_t* out, uint32_t* SFout,
uint32_t* input_offset_by_experts,
@ -233,8 +234,9 @@ void quant_impl(void* output, void* output_scale, void* input,
int const workSizePerRow = k / ELTS_PER_THREAD;
int const totalWorkSize = m_topk * workSizePerRow;
dim3 block(std::min(workSizePerRow, 512));
// Get number of blocks per SM (assume we can fully utilize the SM).
int const numBlocksPerSM = 2048 / block.x;
// Get number of blocks per SM
int const numBlocksPerSM =
vllm_runtime_blocks_per_sm(static_cast<int>(block.x));
dim3 grid(std::min(static_cast<int>((totalWorkSize + block.x - 1) / block.x),
multiProcessorCount * numBlocksPerSM));
while (grid.x <= multiProcessorCount && block.x > 64) {

View File

@ -26,13 +26,14 @@
#include "dispatch_utils.h"
#include "cuda_utils.h"
#include "launch_bounds_utils.h"
#include "nvfp4_utils.cuh"
namespace vllm {
// Use UE4M3 by default.
template <class Type, bool UE8M0_SF = false>
__global__ void __launch_bounds__(512, 4)
__global__ void __launch_bounds__(512, VLLM_BLOCKS_PER_SM(512))
cvt_fp16_to_fp4(int32_t numRows, int32_t numCols, Type const* in,
float const* SFScale, uint32_t* out, uint32_t* SFout) {
using PackedVec = PackedVec<Type>;
@ -75,8 +76,9 @@ void invokeFP4Quantization(int m, int n, T const* input, float const* SFScale,
// Grid, Block size.
// Each thread converts 8 values.
dim3 block(std::min(int(n / ELTS_PER_THREAD), 512));
// Get number of blocks per SM (assume we can fully utilize the SM).
int const numBlocksPerSM = 2048 / block.x;
// Get number of blocks per SM
int const numBlocksPerSM =
vllm_runtime_blocks_per_sm(static_cast<int>(block.x));
dim3 grid(std::min(int(m), multiProcessorCount * numBlocksPerSM));
// Launch the cvt kernel.

View File

@ -576,6 +576,17 @@ __inline__ __device__ Tout scaled_convert(const Tin& x, const float scale) {
TORCH_CHECK(false, \
"Unsupported input type of kv cache: ", SRC_DTYPE); \
} \
} else if (KV_DTYPE == "fp8_ds_mla") { \
if (SRC_DTYPE == at::ScalarType::Float) { \
FN(float, uint8_t, vllm::Fp8KVCacheDataType::kFp8E4M3); \
} else if (SRC_DTYPE == at::ScalarType::Half) { \
FN(uint16_t, uint8_t, vllm::Fp8KVCacheDataType::kFp8E4M3); \
} else if (SRC_DTYPE == at::ScalarType::BFloat16) { \
FN(__nv_bfloat16, uint8_t, vllm::Fp8KVCacheDataType::kFp8E4M3); \
} else { \
TORCH_CHECK(false, \
"Unsupported input type of kv cache: ", SRC_DTYPE); \
} \
} else { \
TORCH_CHECK(false, "Unsupported data type of kv cache: ", KV_DTYPE); \
} \

View File

@ -12,8 +12,8 @@
#include "../vectorization_utils.cuh"
#include "../../dispatch_utils.h"
__device__ __forceinline__ float GroupReduceMax(float val, const int tid) {
unsigned mask = 0xffff;
__device__ __forceinline__ float GroupReduceMax(float val) {
unsigned mask = threadIdx.x % 32 >= 16 ? 0xffff0000 : 0x0000ffff;
val = fmaxf(val, __shfl_xor_sync(mask, val, 8));
val = fmaxf(val, __shfl_xor_sync(mask, val, 4));
@ -86,7 +86,7 @@ __global__ void per_token_group_quant_8bit_kernel(
threads_per_group, // stride in group
scalar_op_cache); // scalar handler
local_absmax = GroupReduceMax(local_absmax, lane_id);
local_absmax = GroupReduceMax(local_absmax);
float y_s = local_absmax / max_8bit;
if constexpr (SCALE_UE8M0) {

View File

@ -25,6 +25,12 @@
#include "../attention/dtype_fp8.cuh"
#include "../quantization/fp8/amd/quant_utils.cuh"
// ROCm 6.2 compatibility: map OCP fp8 types to FNUZ variants if OCP is absent
#if !defined(HIP_FP8_TYPE_OCP)
using __hip_fp8_e4m3 = __hip_fp8_e4m3_fnuz;
using __hip_fp8_e5m2 = __hip_fp8_e5m2_fnuz;
#endif
#if defined(__HIPCC__) && \
(defined(__gfx90a__) || defined(__gfx942__) || defined(__gfx950__))
#define __HIP__GFX9__
@ -34,7 +40,8 @@
#define __HIP__FP8MFMA__
#endif
#if defined(__HIPCC__) && (defined(__gfx1100__) || defined(__gfx1101__))
#if defined(__HIPCC__) && (defined(__gfx1100__) || defined(__gfx1101__) || \
defined(__gfx1150__) || defined(__gfx1151__))
#define __HIP__GFX11__
#endif

View File

@ -5,11 +5,14 @@
torch::Tensor LLMM1(at::Tensor& in_a, at::Tensor& in_b,
const int64_t rows_per_block);
torch::Tensor wvSplitK(at::Tensor& in_a, at::Tensor& in_b,
torch::Tensor wvSplitK(const at::Tensor& in_a, const at::Tensor& in_b,
const std::optional<at::Tensor>& in_bias,
const int64_t CuCount);
void wvSplitKQ(at::Tensor& in_a, at::Tensor& in_b, at::Tensor& out_c,
at::Tensor& scale_a, at::Tensor& scale_b, const int64_t CuCount);
void wvSplitKQ(const at::Tensor& in_a, const at::Tensor& in_b,
const std::optional<at::Tensor>& in_bias, at::Tensor& out_c,
const at::Tensor& scale_a, const at::Tensor& scale_b,
const int64_t CuCount);
void paged_attention(
torch::Tensor& out, torch::Tensor& exp_sums, torch::Tensor& max_logits,

View File

@ -292,8 +292,9 @@ torch::Tensor LLMM1(at::Tensor& in_a, at::Tensor& in_b,
template <typename scalar_t, int THRDS, int YTILE, int WvPrGrp, int A_CHUNK,
int UNRL, int N>
__global__ void __launch_bounds__(WvPrGrp* THRDS)
wvSplitK_hf_sml_(const int K, const int M, const scalar_t* B,
const scalar_t* __restrict__ A, scalar_t* C,
wvSplitK_hf_sml_(const int K, const int M, const int Bx, const int By,
const scalar_t* B, const scalar_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const int _WvPrGrp, const int CuCount) {
constexpr int max_lds_len = LDS_SIZE / 2;
#if defined(__HIP__MI3XX__)
@ -484,7 +485,14 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 63) {
for (int n = 0; n < N; n++) {
for (int i = 0; i < YTILE; i++) {
// if (commitColumn[i]) C[m + i + n * M] = __float2half(sum[n][i]);
if constexpr (std::is_same_v<scalar_t, half>) {
if (BIAS)
sum[n][i] += __half2float(BIAS[(m + i) % Bx + (n % By) * M]);
} else if constexpr (std::is_same_v<scalar_t, __hip_bfloat16>) {
if (BIAS)
sum[n][i] +=
__bfloat162float(BIAS[(m + i) % Bx + (n % By) * M]);
}
C[m + i + n * M] = __float2s<scalar_t>(sum[n][i]);
}
}
@ -529,7 +537,9 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 63) {
for (int n = 0; n < N; n++) {
for (int i = 0; i < YTILE; i++) {
// if (commitColumn[i]) C[n + i + m * N] = __float2half(sum[n][i]);
if (BIAS)
sum4[n][i][0] +=
__bfloat162float(BIAS[(m + i) % Bx + (n % By) * M]);
C[m + i + n * M] = __float2bfloat16(sum4[n][i][0]);
}
}
@ -541,8 +551,10 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
#else // !defined(__HIP__GFX9__) TODO: Add NAVI support
template <typename scalar_t, int THRDS, int YTILE, int WvPrGrp, int A_CHUNK,
int UNRL, int N>
__global__ void wvSplitK_hf_sml_(const int K, const int M, const scalar_t* B,
const scalar_t* __restrict__ A, scalar_t* C,
__global__ void wvSplitK_hf_sml_(const int K, const int M, const int Bx,
const int By, const scalar_t* B,
const scalar_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const int _WvPrGrp, const int CuCount) {
UNREACHABLE_CODE
}
@ -553,8 +565,9 @@ __global__ void wvSplitK_hf_sml_(const int K, const int M, const scalar_t* B,
template <typename scalar_t, int THRDS, int YTILE, int WvPrGrp, int A_CHUNK,
int UNRL, int N>
__global__ void __launch_bounds__(WvPrGrp* THRDS)
wvSplitK_hf_(const int K, const int M, const scalar_t* B,
const scalar_t* __restrict__ A, scalar_t* C,
wvSplitK_hf_(const int K, const int M, const int Bx, const int By,
const scalar_t* B, const scalar_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const int _WvPrGrp, const int CuCount) {
constexpr int max_lds_len = LDS_SIZE / 2;
#if defined(__HIP__MI3XX__)
@ -772,8 +785,17 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 63) {
for (int n = 0; n < N; n++) {
for (int i = 0; i < YTILE; i++) {
if (commitColumn[i])
if (commitColumn[i]) {
if constexpr (std::is_same_v<scalar_t, half>) {
if (BIAS)
sum[n][i] += __half2float(BIAS[(m + i) % Bx + (n % By) * M]);
} else if constexpr (std::is_same_v<scalar_t, __hip_bfloat16>) {
if (BIAS)
sum[n][i] +=
__bfloat162float(BIAS[(m + i) % Bx + (n % By) * M]);
}
C[m + i + n * M] = __float2s<scalar_t>(sum[n][i]);
}
}
}
}
@ -818,8 +840,12 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 63) {
for (int n = 0; n < N; n++) {
for (int i = 0; i < YTILE; i++) {
// if (commitColumn[i]) C[n + i + m * N] = __float2half(sum[n][i]);
C[m + i + n * M] = __float2bfloat16(sum4[n][i][0]);
if (commitColumn[i]) {
if (BIAS)
sum4[n][i][0] +=
__bfloat162float(BIAS[(m + i) % Bx + (n % By) * M]);
C[m + i + n * M] = __float2bfloat16(sum4[n][i][0]);
}
}
}
}
@ -842,8 +868,10 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
#else // !defined(__HIP__GFX9__) TODO: Add NAVI support
template <typename scalar_t, int THRDS, int YTILE, int WvPrGrp, int A_CHUNK,
int UNRL, int N>
__global__ void wvSplitK_hf_(const int K, const int M, const scalar_t* B,
const scalar_t* __restrict__ A, scalar_t* C,
__global__ void wvSplitK_hf_(const int K, const int M, const int Bx,
const int By, const scalar_t* B,
const scalar_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const int _WvPrGrp, const int CuCount) {
UNREACHABLE_CODE
}
@ -854,8 +882,9 @@ __global__ void wvSplitK_hf_(const int K, const int M, const scalar_t* B,
template <typename scalar_t, int THRDS, int YTILE, int WvPrGrp, int A_CHUNK,
int UNRL, int N>
__global__ void __launch_bounds__(WvPrGrp* THRDS)
wvSplitK_hf_big_(const int K, const int M, const scalar_t* B,
const scalar_t* __restrict__ A, scalar_t* C,
wvSplitK_hf_big_(const int K, const int M, const int Bx, const int By,
const scalar_t* B, const scalar_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const int _WvPrGrp, const int CuCount) {
constexpr int max_lds_len = LDS_SIZE / 2;
#if defined(__HIP__MI3XX__)
@ -1124,8 +1153,17 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 63) {
for (int n = 0; n < N; n++) {
for (int i = 0; i < YTILE; i++) {
if (commitColumn[i])
if (commitColumn[i]) {
if constexpr (std::is_same_v<scalar_t, half>) {
if (BIAS)
sum[n][i] += __half2float(BIAS[(m + i) % Bx + (n % By) * M]);
} else if constexpr (std::is_same_v<scalar_t, __hip_bfloat16>) {
if (BIAS)
sum[n][i] +=
__bfloat162float(BIAS[(m + i) % Bx + (n % By) * M]);
}
C[m + i + n * M] = __float2s<scalar_t>(sum[n][i]);
}
}
}
}
@ -1166,8 +1204,12 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 63) {
for (int n = 0; n < N; n++) {
for (int i = 0; i < YTILE; i++) {
// if (commitColumn[i]) C[n + i + m * N] = __float2half(sum[n][i]);
C[m + i + n * M] = __float2bfloat16(sum4[n][i][0]);
if (commitColumn[i]) {
if (BIAS)
sum4[n][i][0] +=
__bfloat162float(BIAS[(m + i) % Bx + (n % By) * M]);
C[m + i + n * M] = __float2bfloat16(sum4[n][i][0]);
}
}
}
}
@ -1190,8 +1232,10 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
#else // !defined(__HIP__GFX9__) TODO: Add NAVI support
template <typename scalar_t, int THRDS, int YTILE, int WvPrGrp, int A_CHUNK,
int UNRL, int N>
__global__ void wvSplitK_hf_big_(const int K, const int M, const scalar_t* B,
const scalar_t* __restrict__ A, scalar_t* C,
__global__ void wvSplitK_hf_big_(const int K, const int M, const int Bx,
const int By, const scalar_t* B,
const scalar_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const int _WvPrGrp, const int CuCount) {
UNREACHABLE_CODE
}
@ -1226,11 +1270,20 @@ int mindiv(int N, int div1, int div2) {
return rtn;
}
torch::Tensor wvSplitK(at::Tensor& in_a, at::Tensor& in_b,
torch::Tensor wvSplitK(const at::Tensor& in_a, const at::Tensor& in_b,
const std::optional<at::Tensor>& in_bias,
const int64_t CuCount) {
auto M_in = in_a.size(0);
auto K_in = in_a.size(1);
auto N_in = in_b.size(0);
auto Bx_in =
(in_bias.has_value() && in_bias->numel() > 0)
? (in_bias->sizes().size() == 2) ? in_bias->size(1) : in_bias->size(0)
: 1;
auto By_in = (in_bias.has_value() && in_bias->numel() > 0 &&
in_bias->sizes().size() == 2)
? in_bias->size(0)
: 1;
TORCH_CHECK(in_a.dtype() == in_b.dtype());
TORCH_CHECK(K_in % 8 == 0, "k % 8 == 0");
@ -1254,18 +1307,18 @@ torch::Tensor wvSplitK(at::Tensor& in_a, at::Tensor& in_b,
if ((K_in * N_in <= max_lds_len) && (M_in % _YTILEs == 0)) { \
int __wvPrGrp = mindiv(M_in, CuCount * _YTILEs, _WvPrGrp); \
wvSplitK_hf_sml_<fptype, 64, _YTILEs, _WvPrGrp, 8, _UNRLs, _N> \
<<<grid, block, 0, stream>>>(K_in, M_in, af4, bf4, c, __wvPrGrp, \
CuCount); \
<<<grid, block, 0, stream>>>(K_in, M_in, Bx_in, By_in, af4, bf4, \
biasf4, c, __wvPrGrp, CuCount); \
} else if (K_in * N_in <= max_lds_len * 1.2) { \
int __wvPrGrp = mindiv(M_in, CuCount * _YTILEm, _WvPrGrp); \
wvSplitK_hf_<fptype, 64, _YTILEm, _WvPrGrp, 8, _UNRLm, _N> \
<<<grid, block, 0, stream>>>(K_in, M_in, af4, bf4, c, __wvPrGrp, \
CuCount); \
<<<grid, block, 0, stream>>>(K_in, M_in, Bx_in, By_in, af4, bf4, \
biasf4, c, __wvPrGrp, CuCount); \
} else { \
int __wvPrGrp = mindiv(M_in, CuCount * _YTILEb, _WvPrGrp); \
wvSplitK_hf_big_<fptype, 64, _YTILEb, _WvPrGrp, 8, _UNRLb, _N> \
<<<grid, block, 0, stream>>>(K_in, M_in, af4, bf4, c, __wvPrGrp, \
CuCount); \
<<<grid, block, 0, stream>>>(K_in, M_in, Bx_in, By_in, af4, bf4, \
biasf4, c, __wvPrGrp, CuCount); \
} \
}
@ -1273,6 +1326,10 @@ torch::Tensor wvSplitK(at::Tensor& in_a, at::Tensor& in_b,
using fptype = typename scalar<scalar_t>::type;
fptype* af4 = reinterpret_cast<fptype*>(in_a.data_ptr());
const fptype* bf4 = reinterpret_cast<const fptype*>(in_b.data_ptr());
const fptype* biasf4 =
(in_bias.has_value() && in_bias->numel() > 0)
? reinterpret_cast<const fptype*>(in_bias->data_ptr())
: nullptr;
fptype* c = reinterpret_cast<fptype*>(out_c.data_ptr());
switch (N_in) {
case 1:
@ -1300,8 +1357,9 @@ torch::Tensor wvSplitK(at::Tensor& in_a, at::Tensor& in_b,
template <typename scalar_t, typename fp8_t, int THRDS, int YTILE, int WvPrGrp,
int A_CHUNK, int UNRL, int N>
__global__ void __launch_bounds__(WvPrGrp* THRDS)
wvSplitKQ_hf_sml_(const int K, const int Kp, const int M, const fp8_t* B,
const fp8_t* __restrict__ A, scalar_t* C,
wvSplitKQ_hf_sml_(const int K, const int Kp, const int M, const int Bx,
const int By, const fp8_t* B, const fp8_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const float* __restrict__ s_A,
const float* __restrict__ s_B, const int _WvPrGrp,
const int CuCount) {
@ -1453,7 +1511,17 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
if (threadIdx.x == 0) {
for (int n = 0; n < N; n++) {
for (int y = 0; y < YTILE; y++) {
C[m + y + n * M] = __float2s<scalar_t>(sum[n][y][0] * sA * sB);
if (y + m >= M) break; // To avoid mem access fault.
sum[n][y][0] *= sA * sB;
if constexpr (std::is_same_v<scalar_t, half>) {
if (BIAS)
sum[n][y][0] += __half2float(BIAS[(m + y) % Bx + (n % By) * M]);
} else if constexpr (std::is_same_v<scalar_t, __hip_bfloat16>) {
if (BIAS)
sum[n][y][0] +=
__bfloat162float(BIAS[(m + y) % Bx + (n % By) * M]);
}
C[m + y + n * M] = __float2s<scalar_t>(sum[n][y][0]); // * sA * sB);
}
}
}
@ -1465,7 +1533,9 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
template <typename scalar_t, typename fp8_t, int THRDS, int YTILE, int WvPrGrp,
int A_CHUNK, int UNRL, int N>
__global__ void wvSplitKQ_hf_sml_(const int K, const int Kp, const int M,
const fp8_t* B, const fp8_t* __restrict__ A,
const int Bx, const int By, const fp8_t* B,
const fp8_t* __restrict__ A,
const scalar_t* __restrict__ BIAS,
scalar_t* C, const float* __restrict__ s_A,
const float* __restrict__ s_B,
const int _WvPrGrp, const int CuCount) {
@ -1477,8 +1547,9 @@ __global__ void wvSplitKQ_hf_sml_(const int K, const int Kp, const int M,
template <typename scalar_t, typename fp8_t, int THRDS, int YTILE, int WvPrGrp,
int A_CHUNK, int UNRL, int N>
__global__ void __launch_bounds__(WvPrGrp* THRDS)
wvSplitKQ_hf_(const int K, const int Kp, const int M, const fp8_t* B,
const fp8_t* __restrict__ A, scalar_t* C,
wvSplitKQ_hf_(const int K, const int Kp, const int M, const int Bx,
const int By, const fp8_t* B, const fp8_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const float* __restrict__ s_A, const float* __restrict__ s_B,
const int _WvPrGrp, const int CuCount) {
constexpr int max_lds_len = LDS_SIZE;
@ -1626,7 +1697,16 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
for (int n = 0; n < N; n++) {
for (int y = 0; y < YTILE; y++) {
if (y + m >= M) break; // To avoid mem access fault.
C[m + y + n * M] = __float2s<scalar_t>(sum[n][y][0] * sA * sB);
sum[n][y][0] *= sA * sB;
if constexpr (std::is_same_v<scalar_t, half>) {
if (BIAS)
sum[n][y][0] += __half2float(BIAS[(m + y) % Bx + (n % By) * M]);
} else if constexpr (std::is_same_v<scalar_t, __hip_bfloat16>) {
if (BIAS)
sum[n][y][0] +=
__bfloat162float(BIAS[(m + y) % Bx + (n % By) * M]);
}
C[m + y + n * M] = __float2s<scalar_t>(sum[n][y][0]);
}
}
}
@ -1638,16 +1718,19 @@ __global__ void __launch_bounds__(WvPrGrp* THRDS)
template <typename scalar_t, typename fp8_t, int THRDS, int YTILE, int WvPrGrp,
int A_CHUNK, int UNRL, int N>
__global__ void wvSplitKQ_hf_(const int K, const int Kp, const int M,
const fp8_t* B, const fp8_t* __restrict__ A,
scalar_t* C, const float* __restrict__ s_A,
const int Bx, const int By, const fp8_t* B,
const fp8_t* __restrict__ A,
const scalar_t* __restrict__ BIAS, scalar_t* C,
const float* __restrict__ s_A,
const float* __restrict__ s_B, const int _WvPrGrp,
const int CuCount) {
UNREACHABLE_CODE
}
#endif // defined(__HIP__MI3XX__) TODO: Add NAVI support
void wvSplitKQ(at::Tensor& in_a, at::Tensor& in_b, at::Tensor& out_c,
at::Tensor& scale_a, at::Tensor& scale_b,
void wvSplitKQ(const at::Tensor& in_a, const at::Tensor& in_b,
const std::optional<at::Tensor>& in_bias, at::Tensor& out_c,
const at::Tensor& scale_a, const at::Tensor& scale_b,
const int64_t CuCount) {
static c10::ScalarType kFp8Type = is_fp8_ocp()
? c10::ScalarType::Float8_e4m3fn
@ -1656,6 +1739,15 @@ void wvSplitKQ(at::Tensor& in_a, at::Tensor& in_b, at::Tensor& out_c,
auto K_in = in_a.size(1);
auto N_in = in_b.size(0);
auto Kp_in = in_a.stride(0);
auto Bx_in =
(in_bias.has_value() && in_bias->numel() > 0)
? (in_bias->sizes().size() == 2) ? in_bias->size(1) : in_bias->size(0)
: 1;
auto By_in = (in_bias.has_value() && in_bias->numel() > 0 &&
in_bias->sizes().size() == 2)
? in_bias->size(0)
: 1;
TORCH_CHECK(K_in % 16 == 0, "k % 16 == 0");
TORCH_CHECK(in_a.dtype() == in_b.dtype() && in_a.dtype() == kFp8Type);
TORCH_CHECK(out_c.dtype() == torch::kFloat16 ||
@ -1673,13 +1765,15 @@ void wvSplitKQ(at::Tensor& in_a, at::Tensor& in_b, at::Tensor& out_c,
if ((K_in * N_in <= max_lds_len) && (M_in % _YTILEs == 0)) { \
int __wvPrGrp = mindiv(M_in, CuCount * _YTILEs, _WvPrGrp); \
wvSplitKQ_hf_sml_<fptype, fp8_t, 64, _YTILEs, _WvPrGrp, 16, _UNRLs, _N> \
<<<grid, block, 0, stream>>>(K_in, Kp_in, M_in, a_ptr, b_ptr, c_ptr, \
s_a, s_b, __wvPrGrp, CuCount); \
<<<grid, block, 0, stream>>>(K_in, Kp_in, M_in, Bx_in, By_in, a_ptr, \
b_ptr, bias_ptr, c_ptr, s_a, s_b, \
__wvPrGrp, CuCount); \
} else { \
int __wvPrGrp = mindiv(M_in, CuCount * _YTILEm, _WvPrGrp); \
wvSplitKQ_hf_<fptype, fp8_t, 64, _YTILEm, _WvPrGrp, 16, _UNRLm, _N> \
<<<grid, block, 0, stream>>>(K_in, Kp_in, M_in, a_ptr, b_ptr, c_ptr, \
s_a, s_b, __wvPrGrp, CuCount); \
<<<grid, block, 0, stream>>>(K_in, Kp_in, M_in, Bx_in, By_in, a_ptr, \
b_ptr, bias_ptr, c_ptr, s_a, s_b, \
__wvPrGrp, CuCount); \
} \
}
@ -1691,6 +1785,9 @@ void wvSplitKQ(at::Tensor& in_a, at::Tensor& in_b, at::Tensor& out_c,
VLLM_DISPATCH_FP8_TYPES(in_a.scalar_type(), "wvSplitKQ", [&] {
auto a_ptr = in_a.data_ptr<fp8_t>();
auto b_ptr = in_b.data_ptr<fp8_t>();
auto bias_ptr = (in_bias.has_value() && in_bias->numel() > 0)
? reinterpret_cast<fptype*>(in_bias->data_ptr())
: nullptr;
switch (N_in) {
case 1:
WVSPLITKQ(16, 2, 2, 2, 2, 2, 2, 1)

View File

@ -22,13 +22,14 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, rocm_ops) {
// Custom gemm op for skinny matrix-matrix multiplication
rocm_ops.def(
"wvSplitK(Tensor in_a, Tensor in_b, int CuCount) -> "
"wvSplitK(Tensor in_a, Tensor in_b, Tensor? in_bias, int CuCount) -> "
"Tensor");
rocm_ops.impl("wvSplitK", torch::kCUDA, &wvSplitK);
// wvSplitK for fp8
rocm_ops.def(
"wvSplitKQ(Tensor in_a, Tensor in_b, Tensor! out_c, Tensor scale_a, "
"wvSplitKQ(Tensor in_a, Tensor in_b, Tensor? in_bias, Tensor! out_c, "
"Tensor scale_a, "
" Tensor scale_b, int CuCount) -> ()");
rocm_ops.impl("wvSplitKQ", torch::kCUDA, &wvSplitKQ);

View File

@ -397,7 +397,7 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
" Tensor a_blockscale, Tensor b_blockscales, Tensor alphas,"
" Tensor problem_sizes, Tensor expert_offsets, Tensor sf_offsets) -> ()",
{stride_tag});
ops.impl("cutlass_fp4_group_mm", torch::kCUDA, &cutlass_fp4_group_mm);
// conditionally compiled so impl registration is in source file
// CUTLASS w8a8 GEMM, supporting symmetric per-tensor or per-row/column
// quantization, as well as bias
@ -713,6 +713,13 @@ TORCH_LIBRARY_EXPAND(CONCAT(TORCH_EXTENSION_NAME, _cache_ops), cache_ops) {
"cp_gather_cache(Tensor src_cache, Tensor! dst, Tensor block_table, "
"Tensor cu_seq_lens, int batch_size, Tensor? seq_starts) -> ()");
cache_ops.impl("cp_gather_cache", torch::kCUDA, &cp_gather_cache);
cache_ops.def(
"indexer_k_quant_and_cache(Tensor k, Tensor! kv_cache, Tensor "
"slot_mapping, "
"int quant_block_size, str kv_cache_dtype) -> ()");
cache_ops.impl("indexer_k_quant_and_cache", torch::kCUDA,
&indexer_k_quant_and_cache);
}
TORCH_LIBRARY_EXPAND(CONCAT(TORCH_EXTENSION_NAME, _cuda_utils), cuda_utils) {

View File

@ -14,6 +14,11 @@ ARG PYTHON_VERSION=3.12
#
# Example:
# docker build --build-arg BUILD_BASE_IMAGE=registry.acme.org/mirror/nvidia/cuda:${CUDA_VERSION}-devel-ubuntu20.04
# Important: We build with an old version of Ubuntu to maintain broad
# compatibility with other Linux OSes. The main reason for this is that the
# glibc version is baked into the distro, and binaries built with one glibc
# version are not backwards compatible with OSes that use an earlier version.
ARG BUILD_BASE_IMAGE=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu20.04
# TODO: Restore to base image after FlashInfer AOT wheel fixed
ARG FINAL_BASE_IMAGE=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu22.04
@ -75,34 +80,19 @@ ARG TARGETPLATFORM
ARG INSTALL_KV_CONNECTORS=false
ENV DEBIAN_FRONTEND=noninteractive
ARG DEADSNAKES_MIRROR_URL
ARG DEADSNAKES_GPGKEY_URL
ARG GET_PIP_URL
# Install Python and other dependencies
# Install system dependencies and uv, then create Python virtual environment
RUN echo 'tzdata tzdata/Areas select America' | debconf-set-selections \
&& echo 'tzdata tzdata/Zones/America select Los_Angeles' | debconf-set-selections \
&& apt-get update -y \
&& apt-get install -y ccache software-properties-common git curl sudo \
&& if [ ! -z ${DEADSNAKES_MIRROR_URL} ] ; then \
if [ ! -z "${DEADSNAKES_GPGKEY_URL}" ] ; then \
mkdir -p -m 0755 /etc/apt/keyrings ; \
curl -L ${DEADSNAKES_GPGKEY_URL} | gpg --dearmor > /etc/apt/keyrings/deadsnakes.gpg ; \
sudo chmod 644 /etc/apt/keyrings/deadsnakes.gpg ; \
echo "deb [signed-by=/etc/apt/keyrings/deadsnakes.gpg] ${DEADSNAKES_MIRROR_URL} $(lsb_release -cs) main" > /etc/apt/sources.list.d/deadsnakes.list ; \
fi ; \
else \
for i in 1 2 3; do \
add-apt-repository -y ppa:deadsnakes/ppa && break || \
{ echo "Attempt $i failed, retrying in 5s..."; sleep 5; }; \
done ; \
fi \
&& apt-get update -y \
&& apt-get install -y python${PYTHON_VERSION} python${PYTHON_VERSION}-dev python${PYTHON_VERSION}-venv \
&& update-alternatives --install /usr/bin/python3 python3 /usr/bin/python${PYTHON_VERSION} 1 \
&& update-alternatives --set python3 /usr/bin/python${PYTHON_VERSION} \
&& ln -sf /usr/bin/python${PYTHON_VERSION}-config /usr/bin/python3-config \
&& curl -sS ${GET_PIP_URL} | python${PYTHON_VERSION} \
&& apt-get install -y ccache software-properties-common git curl sudo python3-pip \
&& curl -LsSf https://astral.sh/uv/install.sh | sh \
&& $HOME/.local/bin/uv venv /opt/venv --python ${PYTHON_VERSION} \
&& rm -f /usr/bin/python3 /usr/bin/python3-config /usr/bin/pip \
&& ln -s /opt/venv/bin/python3 /usr/bin/python3 \
&& ln -s /opt/venv/bin/python3-config /usr/bin/python3-config \
&& ln -s /opt/venv/bin/pip /usr/bin/pip \
&& python3 --version && python3 -m pip --version
ARG PIP_INDEX_URL UV_INDEX_URL
@ -111,9 +101,9 @@ ARG PYTORCH_CUDA_INDEX_BASE_URL
ARG PYTORCH_CUDA_NIGHTLY_INDEX_BASE_URL
ARG PIP_KEYRING_PROVIDER UV_KEYRING_PROVIDER
# Install uv for faster pip installs
RUN --mount=type=cache,target=/root/.cache/uv \
python3 -m pip install uv
# Activate virtual environment and add uv to PATH
ENV PATH="/opt/venv/bin:/root/.local/bin:$PATH"
ENV VIRTUAL_ENV="/opt/venv"
# This timeout (in seconds) is necessary when installing some dependencies via uv since it's likely to time out
# Reference: https://github.com/astral-sh/uv/pull/1694
@ -142,7 +132,7 @@ WORKDIR /workspace
COPY requirements/common.txt requirements/common.txt
COPY requirements/cuda.txt requirements/cuda.txt
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install --system -r requirements/cuda.txt \
uv pip install --python /opt/venv/bin/python3 -r requirements/cuda.txt \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
# cuda arch list used by torch
@ -172,7 +162,7 @@ ENV UV_INDEX_STRATEGY="unsafe-best-match"
ENV UV_LINK_MODE=copy
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install --system -r requirements/build.txt \
uv pip install --python /opt/venv/bin/python3 -r requirements/build.txt \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
COPY . .
@ -269,7 +259,7 @@ COPY requirements/lint.txt requirements/lint.txt
COPY requirements/test.txt requirements/test.txt
COPY requirements/dev.txt requirements/dev.txt
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install --system -r requirements/dev.txt \
uv pip install --python /opt/venv/bin/python3 -r requirements/dev.txt \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
#################### DEV IMAGE ####################
@ -391,19 +381,32 @@ RUN --mount=type=cache,target=/root/.cache/uv bash - <<'BASH'
git clone --depth 1 --recursive --shallow-submodules \
--branch ${FLASHINFER_GIT_REF} \
${FLASHINFER_GIT_REPO} flashinfer
# Exclude CUDA arches for older versions (11.x and 12.0-12.7)
# TODO: Update this to allow setting TORCH_CUDA_ARCH_LIST as a build arg.
if [[ "${CUDA_VERSION}" == 11.* ]]; then
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9"
elif [[ "${CUDA_VERSION}" == 12.[0-7]* ]]; then
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a"
else
# CUDA 12.8+ supports 10.0a and 12.0
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a 10.0a 12.0"
fi
pushd flashinfer
if [ "${FLASHINFER_AOT_COMPILE}" = "true" ]; then
# Exclude CUDA arches for older versions (11.x and 12.0-12.7)
# TODO: Update this to allow setting TORCH_CUDA_ARCH_LIST as a build arg.
if [[ "${CUDA_VERSION}" == 11.* ]]; then
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9"
elif [[ "${CUDA_VERSION}" == 12.[0-7]* ]]; then
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a"
else
# CUDA 12.8+ supports 10.0a and 12.0
FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a 10.0a 12.0"
if [[ "${CUDA_VERSION}" == 12.8.* ]] && [ "$TARGETPLATFORM" = "linux/amd64" ]; then
# NOTE: To make new precompiled wheels, see tools/flashinfer-build.sh
echo "🏗️ Installing FlashInfer from pre-compiled wheel"
uv pip install --system https://wheels.vllm.ai/flashinfer-python/flashinfer_python-0.3.1-cp39-abi3-manylinux1_x86_64.whl \
--extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
if [ "${FLASHINFER_AOT_COMPILE}" = "true" ]; then
# Download pre-compiled cubins
TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
python3 -m flashinfer --download-cubin || echo "WARNING: Failed to download flashinfer cubins."
fi
elif [ "${FLASHINFER_AOT_COMPILE}" = "true" ]; then
echo "🏗️ Installing FlashInfer with AOT compilation for arches: ${FI_TORCH_CUDA_ARCH_LIST}"
export FLASHINFER_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}"
# HACK: We need these to run flashinfer.aot before installing flashinfer, get from the package in the future
uv pip install --system cuda-python==$(echo $CUDA_VERSION | cut -d. -f1,2) pynvml==$(echo $CUDA_VERSION | cut -d. -f1) nvidia-nvshmem-cu$(echo $CUDA_VERSION | cut -d. -f1)
# Build AOT kernels
TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
python3 -m flashinfer.aot
@ -443,7 +446,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
ARG DEEPGEMM_GIT_REF
COPY tools/install_deepgemm.sh /tmp/install_deepgemm.sh
RUN --mount=type=cache,target=/root/.cache/uv \
VLLM_DOCKER_BUILD_CONTEXT=1 /tmp/install_deepgemm.sh --cuda-version "${CUDA_VERSION}" ${DEEPGEMM_GIT_REF:+--ref "$DEEPGEMM_GIT_REF"}
VLLM_DOCKER_BUILD_CONTEXT=1 TORCH_CUDA_ARCH_LIST="9.0a 10.0a" /tmp/install_deepgemm.sh --cuda-version "${CUDA_VERSION}" ${DEEPGEMM_GIT_REF:+--ref "$DEEPGEMM_GIT_REF"}
COPY tools/install_gdrcopy.sh install_gdrcopy.sh
RUN set -eux; \
@ -461,6 +464,12 @@ ENV CUDA_HOME=/usr/local/cuda
RUN export TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST:-9.0a+PTX}" \
&& bash install_python_libraries.sh
# CUDA image changed from /usr/local/nvidia to /usr/local/cuda in 12.8 but will
# return to /usr/local/nvidia in 13.0 to allow container providers to mount drivers
# consistently from the host (see https://github.com/vllm-project/vllm/issues/18859).
# Until then, add /usr/local/nvidia/lib64 before the image cuda path to allow override.
ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib64:${LD_LIBRARY_PATH}
#################### vLLM installation IMAGE ####################
#################### TEST IMAGE ####################
@ -533,7 +542,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
else \
BITSANDBYTES_VERSION="0.46.1"; \
fi; \
uv pip install --system accelerate hf_transfer modelscope "bitsandbytes>=${BITSANDBYTES_VERSION}" 'timm>=1.0.17' boto3 runai-model-streamer runai-model-streamer[s3]
uv pip install --system accelerate hf_transfer modelscope "bitsandbytes>=${BITSANDBYTES_VERSION}" 'timm>=1.0.17' 'runai-model-streamer[s3]>=0.14.0'
ENV VLLM_USAGE_SOURCE production-docker-image
@ -546,5 +555,5 @@ ENTRYPOINT ["./sagemaker-entrypoint.sh"]
FROM vllm-openai-base AS vllm-openai
ENTRYPOINT ["python3", "-m", "vllm.entrypoints.openai.api_server"]
ENTRYPOINT ["vllm", "serve"]
#################### OPENAI API SERVER ####################

View File

@ -47,7 +47,7 @@ ENV PATH="$VIRTUAL_ENV/bin:$PATH"
ENV UV_HTTP_TIMEOUT=500
# Install Python dependencies
# Install Python dependencies
ENV PIP_EXTRA_INDEX_URL=${PIP_EXTRA_INDEX_URL}
ENV UV_EXTRA_INDEX_URL=${PIP_EXTRA_INDEX_URL}
ENV UV_INDEX_STRATEGY="unsafe-best-match"
@ -104,7 +104,7 @@ RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=cache,target=/root/.cache/ccache \
--mount=type=cache,target=/workspace/vllm/.deps,sharing=locked \
--mount=type=bind,source=.git,target=.git \
VLLM_TARGET_DEVICE=cpu python3 setup.py bdist_wheel
VLLM_TARGET_DEVICE=cpu python3 setup.py bdist_wheel
######################### TEST DEPS #########################
FROM base AS vllm-test-deps
@ -114,13 +114,10 @@ WORKDIR /workspace/vllm
RUN --mount=type=bind,src=requirements/test.in,target=requirements/test.in \
cp requirements/test.in requirements/cpu-test.in && \
sed -i '/mamba_ssm/d' requirements/cpu-test.in && \
sed -i 's/^torch==.*/torch==2.6.0/g' requirements/cpu-test.in && \
sed -i 's/torchaudio.*/torchaudio/g' requirements/cpu-test.in && \
sed -i 's/torchvision.*/torchvision/g' requirements/cpu-test.in && \
uv pip compile requirements/cpu-test.in -o requirements/cpu-test.txt --index-strategy unsafe-best-match --torch-backend cpu
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install -r requirements/cpu-test.txt
uv pip install -r requirements/cpu-test.txt
######################### DEV IMAGE #########################
FROM vllm-build AS vllm-dev
@ -133,12 +130,12 @@ RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
# install development dependencies (for testing)
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install -e tests/vllm_test_utils
uv pip install -e tests/vllm_test_utils
RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=cache,target=/root/.cache/ccache \
--mount=type=bind,source=.git,target=.git \
VLLM_TARGET_DEVICE=cpu python3 setup.py develop
VLLM_TARGET_DEVICE=cpu python3 setup.py develop
COPY --from=vllm-test-deps /workspace/vllm/requirements/cpu-test.txt requirements/test.txt
@ -163,11 +160,12 @@ ADD ./benchmarks/ ./benchmarks/
ADD ./vllm/collect_env.py .
ADD ./.buildkite/ ./.buildkite/
# Create symlink for vllm-workspace to maintain CI compatibility
RUN ln -sf /workspace /vllm-workspace
# install development dependencies (for testing)
RUN --mount=type=cache,target=/root/.cache/uv \
uv pip install -e tests/vllm_test_utils
ENTRYPOINT ["bash"]
uv pip install -e tests/vllm_test_utils
######################### RELEASE IMAGE #########################
FROM base AS vllm-openai
@ -179,4 +177,4 @@ RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=bind,from=vllm-build,src=/workspace/vllm/dist,target=dist \
uv pip install dist/*.whl
ENTRYPOINT ["python3", "-m", "vllm.entrypoints.openai.api_server"]
ENTRYPOINT ["vllm", "serve"]

View File

@ -6,7 +6,7 @@ ARG CUDA_VERSION=12.8.0
#
#################### BASE BUILD IMAGE ####################
# prepare basic build environment
FROM nvidia/cuda:${CUDA_VERSION}-devel-ubuntu20.04 AS base
FROM nvidia/cuda:${CUDA_VERSION}-devel-ubuntu22.04 AS base
ARG CUDA_VERSION=12.8.0
ARG PYTHON_VERSION=3.12
ARG TARGETPLATFORM

View File

@ -314,4 +314,4 @@ WORKDIR /workspace/
RUN ln -s /workspace/vllm/tests && ln -s /workspace/vllm/examples && ln -s /workspace/vllm/benchmarks
ENTRYPOINT ["python", "-m", "vllm.entrypoints.openai.api_server"]
ENTRYPOINT ["vllm", "serve"]

View File

@ -15,7 +15,7 @@ FROM ${BASE_IMAGE} AS base
ENV PATH=/opt/rocm/llvm/bin:/opt/rocm/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
ENV ROCM_PATH=/opt/rocm
ENV LD_LIBRARY_PATH=/opt/rocm/lib:/usr/local/lib:
ARG PYTORCH_ROCM_ARCH=gfx90a;gfx942;gfx950;gfx1100;gfx1101;gfx1200;gfx1201
ARG PYTORCH_ROCM_ARCH=gfx90a;gfx942;gfx950;gfx1100;gfx1101;gfx1200;gfx1201;gfx1150;gfx1151
ENV PYTORCH_ROCM_ARCH=${PYTORCH_ROCM_ARCH}
ENV AITER_ROCM_ARCH=gfx942;gfx950
@ -65,8 +65,6 @@ ARG PYTORCH_BRANCH
ARG PYTORCH_VISION_BRANCH
ARG PYTORCH_REPO
ARG PYTORCH_VISION_REPO
ARG FA_BRANCH
ARG FA_REPO
RUN git clone ${PYTORCH_REPO} pytorch
RUN cd pytorch && git checkout ${PYTORCH_BRANCH} && \
pip install -r requirements.txt && git submodule update --init --recursive \
@ -77,14 +75,20 @@ RUN git clone ${PYTORCH_VISION_REPO} vision
RUN cd vision && git checkout ${PYTORCH_VISION_BRANCH} \
&& python3 setup.py bdist_wheel --dist-dir=dist \
&& pip install dist/*.whl
RUN mkdir -p /app/install && cp /app/pytorch/dist/*.whl /app/install \
&& cp /app/vision/dist/*.whl /app/install
FROM base AS build_fa
ARG FA_BRANCH
ARG FA_REPO
RUN --mount=type=bind,from=build_pytorch,src=/app/install/,target=/install \
pip install /install/*.whl
RUN git clone ${FA_REPO}
RUN cd flash-attention \
&& git checkout ${FA_BRANCH} \
&& git submodule update --init \
&& GPU_ARCHS=$(echo ${PYTORCH_ROCM_ARCH} | sed -e 's/;gfx1[0-9]\{3\}//g') python3 setup.py bdist_wheel --dist-dir=dist
RUN mkdir -p /app/install && cp /app/pytorch/dist/*.whl /app/install \
&& cp /app/vision/dist/*.whl /app/install \
&& cp /app/flash-attention/dist/*.whl /app/install
RUN mkdir -p /app/install && cp /app/flash-attention/dist/*.whl /app/install
FROM base AS build_aiter
ARG AITER_BRANCH
@ -103,6 +107,8 @@ FROM base AS debs
RUN mkdir /app/debs
RUN --mount=type=bind,from=build_triton,src=/app/install/,target=/install \
cp /install/*.whl /app/debs
RUN --mount=type=bind,from=build_fa,src=/app/install/,target=/install \
cp /install/*.whl /app/debs
RUN --mount=type=bind,from=build_amdsmi,src=/app/install/,target=/install \
cp /install/*.whl /app/debs
RUN --mount=type=bind,from=build_pytorch,src=/app/install/,target=/install \
@ -111,13 +117,7 @@ RUN --mount=type=bind,from=build_aiter,src=/app/install/,target=/install \
cp /install/*.whl /app/debs
FROM base AS final
RUN --mount=type=bind,from=build_triton,src=/app/install/,target=/install \
pip install /install/*.whl
RUN --mount=type=bind,from=build_amdsmi,src=/app/install/,target=/install \
pip install /install/*.whl
RUN --mount=type=bind,from=build_pytorch,src=/app/install/,target=/install \
pip install /install/*.whl
RUN --mount=type=bind,from=build_aiter,src=/app/install/,target=/install \
RUN --mount=type=bind,from=debs,src=/app/debs,target=/install \
pip install /install/*.whl
ARG BASE_IMAGE
@ -141,4 +141,4 @@ RUN echo "BASE_IMAGE: ${BASE_IMAGE}" > /app/versions.txt \
&& echo "FA_BRANCH: ${FA_BRANCH}" >> /app/versions.txt \
&& echo "FA_REPO: ${FA_REPO}" >> /app/versions.txt \
&& echo "AITER_BRANCH: ${AITER_BRANCH}" >> /app/versions.txt \
&& echo "AITER_REPO: ${AITER_REPO}" >> /app/versions.txt
&& echo "AITER_REPO: ${AITER_REPO}" >> /app/versions.txt

View File

@ -309,4 +309,4 @@ USER 2000
WORKDIR /home/vllm
# Set the default entrypoint
ENTRYPOINT ["python", "-m", "vllm.entrypoints.openai.api_server"]
ENTRYPOINT ["vllm", "serve"]

View File

@ -69,4 +69,4 @@ RUN --mount=type=cache,target=/root/.cache/pip \
# install development dependencies (for testing)
RUN python3 -m pip install -e tests/vllm_test_utils
ENTRYPOINT ["python3", "-m", "vllm.entrypoints.openai.api_server"]
ENTRYPOINT ["vllm", "serve"]

View File

@ -1,2 +1,2 @@
search:
boost: 0.5
exclude: true

Binary file not shown.

After

Width:  |  Height:  |  Size: 627 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 350 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 814 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 267 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 354 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 781 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 51 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 359 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 82 KiB

View File

@ -2,6 +2,7 @@
We host regular meetups in San Francisco Bay Area every 2 months. We will share the project updates from the vLLM team and have guest speakers from the industry to share their experience and insights. Please find the materials of our previous meetups below:
- [vLLM Toronto Meetup](https://luma.com/e80e0ymm), September 25th 2025. [[Slides]](https://docs.google.com/presentation/d/1IYJYmJcu9fLpID5N5RbW_vO0XLo0CGOR14IXOjB61V8/edit?usp=sharing)
- [vLLM Shenzhen Meetup](https://mp.weixin.qq.com/s/k8ZBO1u2_2odgiKWH_GVTQ), August 30th 2025. [[Slides]](https://drive.google.com/drive/folders/1Ua2SVKVSu-wp5vou_6ElraDt2bnKhiEA)
- [vLLM Singapore Meetup](https://www.sginnovate.com/event/vllm-sg-meet), August 27th 2025. [[Slides]](https://drive.google.com/drive/folders/1ncf3GyqLdqFaB6IeB834E5TZJPLAOiXZ?usp=sharing)
- [vLLM Shanghai Meetup](https://mp.weixin.qq.com/s/pDmAXHcN7Iqc8sUKgJgGtg), August 23rd 2025. [[Slides]](https://drive.google.com/drive/folders/1OvLx39wnCGy_WKq8SiVKf7YcxxYI3WCH)

View File

@ -139,9 +139,9 @@ there is relatively little gain from TP. On the other hand, TP incurs significan
overhead because of all-reduce being performed after every layer.
Given this, it may be advantageous to instead shard the batched input data using TP, essentially
performing batch-level DP. This has been shown to improve the throughput by around 10% for
performing batch-level DP. This has been shown to improve the throughput and TTFT by around 10% for
`tensor_parallel_size=8`. For vision encoders that use hardware-unoptimized Conv3D operations,
batch-level DP can provide another 40% increase to throughput compared to regular TP.
batch-level DP can provide another 40% improvement compared to regular TP.
Nevertheless, since the weights of the multi-modal encoder are replicated across each TP rank,
there will be a minor increase in memory consumption and may cause OOM if you can barely fit the model already.
@ -172,14 +172,15 @@ Batch-level DP needs to be implemented on a per-model basis,
and enabled by setting `supports_encoder_tp_data = True` in the model class.
Regardless, you need to set `mm_encoder_tp_mode="data"` in engine arguments to use this feature.
Known supported models:
Known supported models (with corresponding benchmarks):
- GLM-4.5V GLM-4.1V (<gh-pr:23168>)
- dots_ocr (<gh-pr:25466>)
- GLM-4.1V or above (<gh-pr:23168>)
- InternVL (<gh-pr:23909>)
- Kimi-VL (<gh-pr:23817>)
- Llama4 (<gh-pr:18368>)
- MiniCPM-V-2.5 or above (<gh-pr:23327>, <gh-pr:23948>)
- Qwen2.5-VL (<gh-pr:22742>)
- Qwen2-VL or above (<gh-pr:22742>, <gh-pr:24955>, <gh-pr:25445>)
- Step3 (<gh-pr:22697>)
## Input Processing

View File

@ -661,8 +661,7 @@ Benchmark the performance of multi-modal requests in vLLM.
Start vLLM:
```bash
python -m vllm.entrypoints.openai.api_server \
--model Qwen/Qwen2.5-VL-7B-Instruct \
vllm serve Qwen/Qwen2.5-VL-7B-Instruct \
--dtype bfloat16 \
--limit-mm-per-prompt '{"image": 1}' \
--allowed-local-media-path /path/to/sharegpt4v/images
@ -680,7 +679,7 @@ vllm bench serve \
--save-result \
--result-dir ~/vllm_benchmark_results \
--save-detailed \
--endpoint /v1/chat/completion
--endpoint /v1/chat/completions
```
##### Videos (ShareGPT4Video)
@ -688,8 +687,7 @@ vllm bench serve \
Start vLLM:
```bash
python -m vllm.entrypoints.openai.api_server \
--model Qwen/Qwen2.5-VL-7B-Instruct \
vllm serve Qwen/Qwen2.5-VL-7B-Instruct \
--dtype bfloat16 \
--limit-mm-per-prompt '{"video": 1}' \
--allowed-local-media-path /path/to/sharegpt4video/videos
@ -707,7 +705,7 @@ vllm bench serve \
--save-result \
--result-dir ~/vllm_benchmark_results \
--save-detailed \
--endpoint /v1/chat/completion
--endpoint /v1/chat/completions
```
##### Synthetic Random Images (random-mm)
@ -823,6 +821,30 @@ The latest performance results are hosted on the public [vLLM Performance Dashbo
More information on the performance benchmarks and their parameters can be found in [Benchmark README](https://github.com/intel-ai-tce/vllm/blob/more_cpu_models/.buildkite/nightly-benchmarks/README.md) and [performance benchmark description](gh-file:.buildkite/nightly-benchmarks/performance-benchmarks-descriptions.md).
### Continuous Benchmarking
The continuous benchmarking provides automated performance monitoring for vLLM across different models and GPU devices. This helps track vLLM's performance characteristics over time and identify any performance regressions or improvements.
#### How It Works
The continuous benchmarking is triggered via a [GitHub workflow CI](https://github.com/pytorch/pytorch-integration-testing/actions/workflows/vllm-benchmark.yml) in the PyTorch infrastructure repository, which runs automatically every 4 hours. The workflow executes three types of performance tests:
- **Serving tests**: Measure request handling and API performance
- **Throughput tests**: Evaluate token generation rates
- **Latency tests**: Assess response time characteristics
#### Benchmark Configuration
The benchmarking currently runs on a predefined set of models configured in the [vllm-benchmarks directory](https://github.com/pytorch/pytorch-integration-testing/tree/main/vllm-benchmarks/benchmarks). To add new models for benchmarking:
1. Navigate to the appropriate GPU directory in the benchmarks configuration
2. Add your model specifications to the corresponding configuration files
3. The new models will be included in the next scheduled benchmark run
#### Viewing Results
All continuous benchmarking results are automatically published to the public [vLLM Performance Dashboard](https://hud.pytorch.org/benchmark/llms?repoName=vllm-project%2Fvllm).
[](){ #nightly-benchmarks }
## Nightly Benchmarks

View File

@ -66,35 +66,12 @@ Further update the model as follows:
!!! important
The returned `multimodal_embeddings` must be either a **3D [torch.Tensor][]** of shape `(num_items, feature_size, hidden_size)`, or a **list / tuple of 2D [torch.Tensor][]'s** of shape `(feature_size, hidden_size)`, so that `multimodal_embeddings[i]` retrieves the embeddings generated from the `i`-th multimodal data item (e.g, image) of the request.
- Implement [get_input_embeddings][vllm.model_executor.models.interfaces.SupportsMultiModal.get_input_embeddings] to merge `multimodal_embeddings` with text embeddings from the `input_ids`. If input processing for the model is implemented correctly (see sections below), then you can leverage the utility function we provide to easily merge the embeddings.
!!! note
By default, vLLM merges the multimodal embeddings into text embeddings depending on the information of their locations defined in
[PlaceholderRange][vllm.multimodal.inputs.PlaceholderRange] from input processing.
This logic can be found at [get_input_embeddings][vllm.model_executor.models.interfaces.SupportsMultiModal.get_input_embeddings].
??? code
```python
from .utils import merge_multimodal_embeddings
class YourModelForImage2Seq(nn.Module):
...
def get_input_embeddings(
self,
input_ids: torch.Tensor,
multimodal_embeddings: Optional[MultiModalEmbeddings] = None,
) -> torch.Tensor:
# `get_input_embeddings` should already be implemented for the language
# model as one of the requirements of basic vLLM model implementation.
inputs_embeds = self.language_model.get_input_embeddings(input_ids)
if multimodal_embeddings is not None:
inputs_embeds = merge_multimodal_embeddings(
input_ids=input_ids,
inputs_embeds=inputs_embeds,
multimodal_embeddings=multimodal_embeddings,
placeholder_token_id=self.config.image_token_index)
return inputs_embeds
```
You may override this method if additional logic is required for your model when merging embeddings.
- Implement [get_language_model][vllm.model_executor.models.interfaces.SupportsMultiModal.get_language_model] getter to provide stable access to the underlying language model.
@ -281,17 +258,21 @@ Assuming that the memory usage increases with the number of tokens, the dummy in
self,
seq_len: int,
mm_counts: Mapping[str, int],
mm_options: Optional[Mapping[str, BaseDummyOptions]] = None,
) -> MultiModalDataDict:
num_images = mm_counts.get("image", 0)
target_width, target_height = \
self.info.get_image_size_with_most_features()
image_overrides = mm_options.get("image") if mm_options else None
return {
"image":
self._get_dummy_images(width=target_width,
height=target_height,
num_images=num_images)
num_images=num_images,
overrides=image_overrides)
}
```
@ -461,16 +442,20 @@ Assuming that the memory usage increases with the number of tokens, the dummy in
self,
seq_len: int,
mm_counts: Mapping[str, int],
mm_options: Optional[Mapping[str, BaseDummyOptions]] = None,
) -> MultiModalDataDict:
target_width, target_height = \
self.info.get_image_size_with_most_features()
num_images = mm_counts.get("image", 0)
image_overrides = mm_options.get("image") if mm_options else None
return {
"image":
self._get_dummy_images(width=target_width,
height=target_height,
num_images=num_images)
num_images=num_images,
overrides=image_overrides)
}
```

View File

@ -39,8 +39,7 @@ Refer to <gh-file:examples/offline_inference/simple_profiling.py> for an example
```bash
VLLM_TORCH_PROFILER_DIR=./vllm_profile \
python -m vllm.entrypoints.openai.api_server \
--model meta-llama/Meta-Llama-3-70B
vllm serve meta-llama/Meta-Llama-3-70B
```
vllm bench command:
@ -160,6 +159,22 @@ GUI example:
<img width="1799" alt="Screenshot 2025-03-05 at 11 48 42AM" src="https://github.com/user-attachments/assets/c7cff1ae-6d6f-477d-a342-bd13c4fc424c" />
## Continuous Profiling
There is a [GitHub CI workflow](https://github.com/pytorch/pytorch-integration-testing/actions/workflows/vllm-profiling.yml) in the PyTorch infrastructure repository that provides continuous profiling for different models on vLLM. This automated profiling helps track performance characteristics over time and across different model configurations.
### How It Works
The workflow currently runs weekly profiling sessions for selected models, generating detailed performance traces that can be analyzed using different tools to identify performance regressions or optimization opportunities. But, it can be triggered manually as well, using the Github Action tool.
### Adding New Models
To extend the continuous profiling to additional models, you can modify the [profiling-tests.json](https://github.com/pytorch/pytorch-integration-testing/blob/main/vllm-profiling/cuda/profiling-tests.json) configuration file in the PyTorch integration testing repository. Simply add your model specifications to this file to include them in the automated profiling runs.
### Viewing Profiling Results
The profiling traces generated by the continuous profiling workflow are publicly available on the [vLLM Performance Dashboard](https://hud.pytorch.org/benchmark/llms?repoName=vllm-project%2Fvllm). Look for the **Profiling traces** table to access and download the traces for different models and runs.
## Profiling vLLM Python Code
The Python standard library includes
@ -208,3 +223,11 @@ One example is [snakeviz](https://jiffyclub.github.io/snakeviz/).
pip install snakeviz
snakeviz expensive_function.prof
```
### Analyzing Garbage Collection Costs
Leverage VLLM_GC_DEBUG environment variable to debug GC costs.
- VLLM_GC_DEBUG=1: enable GC debugger with gc.collect elpased times
- VLLM_GC_DEBUG='{"top_objects":5}': enable GC debugger to log top 5
collected objects for each gc.collect

View File

@ -19,8 +19,7 @@ pip install -U "autogen-agentchat" "autogen-ext[openai]"
1. Start the vLLM server with the supported chat completion model, e.g.
```bash
python -m vllm.entrypoints.openai.api_server \
--model mistralai/Mistral-7B-Instruct-v0.2
vllm serve mistralai/Mistral-7B-Instruct-v0.2
```
1. Call it with AutoGen:

View File

@ -0,0 +1,170 @@
# Hugging Face Inference Endpoints
## Overview
Models compatible with vLLM can be deployed on Hugging Face Inference Endpoints, either starting from the [Hugging Face Hub](https://huggingface.co) or directly from the [Inference Endpoints](https://endpoints.huggingface.co/) interface. This allows you to serve models in a fully managed environment with GPU acceleration, auto-scaling, and monitoring, without managing the infrastructure manually.
For advanced details on vLLM integration and deployment options, see [Advanced Deployment Details](#advanced-deployment-details).
## Deployment Methods
- [**Method 1: Deploy from the Catalog.**](#method-1-deploy-from-the-catalog) One-click deploy models from the Hugging Face Hub with ready-made optimized configurations.
- [**Method 2: Guided Deployment (Transformers Models).**](#method-2-guided-deployment-transformers-models) Instantly deploy models tagged with `transformers` from the Hub UI using the **Deploy** button.
- [**Method 3: Manual Deployment (Advanced Models).**](#method-3-manual-deployment-advanced-models) For models that either use custom code with the `transformers` tag, or dont run with standard `transformers` but are supported by vLLM. This method requires manual configuration.
### Method 1: Deploy from the Catalog
This is the easiest way to get started with vLLM on Hugging Face Inference Endpoints. You can browse a catalog of models with verified and optimized deployment configuration at [Inference Endpoints](https://endpoints.huggingface.co/catalog) to maximize performance.
1. Go to [Endpoints Catalog](https://endpoints.huggingface.co/catalog) and in the **Inference Server** options, select `vLLM`.This will display the current list of models with optimized preconfigured options.
![Endpoints Catalog](../../assets/deployment/hf-inference-endpoints-catalog.png)
1. Select the desired model and click **Create Endpoint**.
![Create Endpoint](../../assets/deployment/hf-inference-endpoints-create-endpoint.png)
1. Once the deployment is ready, you can use the endpoint. Update the `DEPLOYMENT_URL` with the URL provided in the console, remembering to append `/v1` as required.
```python
# pip install openai
from openai import OpenAI
import os
client = OpenAI(
base_url = DEPLOYMENT_URL,
api_key = os.environ["HF_TOKEN"] # https://huggingface.co/settings/tokens
)
chat_completion = client.chat.completions.create(
model = "HuggingFaceTB/SmolLM3-3B",
messages = [
{
"role": "user",
"content": [
{
"type": "text",
"text": "Give me a brief explanation of gravity in simple terms."
}
]
}
],
stream = True
)
for message in chat_completion:
print(message.choices[0].delta.content, end = "")
```
!!! note
The catalog provides models optimized for vLLM, including GPU settings and inference engine configurations. You can monitor the endpoint and update the **container or its configuration** from the Inference Endpoints UI.
### Method 2: Guided Deployment (Transformers Models)
This method applies to models with the `transformers` library tag in their metadata. It allows you to deploy a model directly from the Hub UI without manual configuration.
1. Navigate to a model on [Hugging Face Hub](https://huggingface.co/models).
For this example we will use the [`ibm-granite/granite-docling-258M`](https://huggingface.co/ibm-granite/granite-docling-258M) model. You can verify that the model is compatible by checking the front matter in the [README](https://huggingface.co/ibm-granite/granite-docling-258M/blob/main/README.md), where the library is tagged as `library: transformers`.
2. Locate the **Deploy** button. The button appears for models tagged with `transformers` at the top right of the [model card](https://huggingface.co/ibm-granite/granite-docling-258M).
![Locate deploy button](../../assets/deployment/hf-inference-endpoints-locate-deploy-button.png)
3. Click to **Deploy** button > **HF Inference Endpoints**. You will be taken to the Inference Endpoints interface to configure the deployment.
![Click deploy button](../../assets/deployment/hf-inference-endpoints-click-deploy-button.png)
4. Select the Hardware (we choose AWS>GPU>T4 for the example) and Container Configuration. Choose `vLLM` as the container type and finalize the deployment pressing **Create Endpoint**.
![Select Hardware](../../assets/deployment/hf-inference-endpoints-select-hardware.png)
5. Use the deployed endpoint. Update the `DEPLOYMENT_URL` with the URL provided in the console (remember to add `/v1` needed). You can then use your endpoint programmatically or via the SDK.
```python
# pip install openai
from openai import OpenAI
import os
client = OpenAI(
base_url = DEPLOYMENT_URL,
api_key = os.environ["HF_TOKEN"] # https://huggingface.co/settings/tokens
)
chat_completion = client.chat.completions.create(
model = "ibm-granite/granite-docling-258M",
messages = [
{
"role": "user",
"content": [
{
"type": "image_url",
"image_url": {
"url": "https://huggingface.co/ibm-granite/granite-docling-258M/resolve/main/assets/new_arxiv.png"
}
},
{
"type": "text",
"text": "Convert this page to docling."
}
]
}
],
stream = True
)
for message in chat_completion:
print(message.choices[0].delta.content, end = "")
```
!!! note
This method uses best-guess defaults. You may need to adjust the configuration to fit your specific requirements.
### Method 3: Manual Deployment (Advanced Models)
Some models require manual deployment because they:
- Use custom code with the `transformers` tag
- Don't run with standard `transformers` but are supported by `vLLM`
These models cannot be deployed using the **Deploy** button on the model card.
In this guide, we demonstrate manual deployment using the [rednote-hilab/dots.ocr](https://huggingface.co/rednote-hilab/dots.ocr) model, an OCR model integrated with vLLM (see vLLM [PR](https://github.com/vllm-project/vllm/pull/24645)).
1. Start a new deployment. Go to [Inference Endpoints](https://endpoints.huggingface.co/) and click `New`.
![New Endpoint](../../assets/deployment/hf-inference-endpoints-new-endpoint.png)
2. Search the model in the Hub. In the dialog, switch to **Hub** and search for the desired model.
![Select model](../../assets/deployment/hf-inference-endpoints-select-model.png)
3. Choosing infrastructure. On the configuration page, select the cloud provider and hardware from the available options.
For this demo, we choose AWS and L4 GPU. Adjust according to your hardware needs.
![Choose Infra](../../assets/deployment/hf-inference-endpoints-choose-infra.png)
4. Configure the container. Scroll to the **Container Configuration** and select `vLLM` as the container type.
![Configure Container](../../assets/deployment/hf-inference-endpoints-configure-container.png)
5. Create the endpoint. Click **Create Endpoint** to deploy the model.
Once the endpoint is ready, you can use it with the OpenAI Completion API, cURL, or other SDKs. Remember to append `/v1` to the deployment URL if needed.
!!! note
You can adjust the **container settings** (Container URI, Container Arguments) from the Inference Endpoints UI and press **Update Endpoint**. This redeploys the endpoint with the updated container configuration. Changes to the model itself require creating a new endpoint or redeploying with a different model. For example, for this demo, you may need to update the Container URI to the nightly image (`vllm/vllm-openai:nightly`) and add the `--trust-remote-code` flag in the container arguments.
## Advanced Deployment Details
With the [transformers backend integration](https://blog.vllm.ai/2025/04/11/transformers-backend.html), vLLM now offers Day 0 support for any model compatible with `transformers`. This means you can deploy such models immediately, leveraging vLLMs optimized inference without additional backend modifications.
Hugging Face Inference Endpoints provides a fully managed environment for serving models via vLLM. You can deploy models without configuring servers, installing dependencies, or managing clusters. Endpoints also support deployment across multiple cloud providers (AWS, Azure, GCP) without the need for separate accounts.
The platform integrates seamlessly with the Hugging Face Hub, allowing you to deploy any vLLM- or `transformers`-compatible model, track usage, and update the inference engine directly. The vLLM engine comes preconfigured, enabling optimized inference and easy switching between models or engines without modifying your code. This setup simplifies production deployment: endpoints are ready in minutes, include monitoring and logging, and let you focus on serving models rather than maintaining infrastructure.
## Next Steps
- Explore the [Inference Endpoints](https://endpoints.huggingface.co/catalog) model catalog
- Read the Inference Endpoints [documentation](https://huggingface.co/docs/inference-endpoints/en/index)
- Learn about [Inference Endpoints engines](https://huggingface.co/docs/inference-endpoints/en/engines/vllm)
- Understand the [transformers backend integration](https://blog.vllm.ai/2025/04/11/transformers-backend.html)

View File

@ -20,7 +20,7 @@ To get started with Open WebUI using vLLM, follow these steps:
For example:
```console
python -m vllm.entrypoints.openai.api_server --host 0.0.0.0 --port 8000
vllm serve <model> --host 0.0.0.0 --port 8000
```
3. Start the Open WebUI Docker container:

View File

@ -32,6 +32,7 @@ See the vLLM SkyPilot YAML for serving, [serving.yaml](https://github.com/skypil
ports: 8081 # Expose to internet traffic.
envs:
PYTHONUNBUFFERED: 1
MODEL_NAME: meta-llama/Meta-Llama-3-8B-Instruct
HF_TOKEN: <your-huggingface-token> # Change to your own huggingface token, or use --env to pass.
@ -47,9 +48,8 @@ See the vLLM SkyPilot YAML for serving, [serving.yaml](https://github.com/skypil
run: |
conda activate vllm
echo 'Starting vllm api server...'
python -u -m vllm.entrypoints.openai.api_server \
vllm serve $MODEL_NAME \
--port 8081 \
--model $MODEL_NAME \
--trust-remote-code \
--tensor-parallel-size $SKYPILOT_NUM_GPUS_PER_NODE \
2>&1 | tee api_server.log &
@ -131,6 +131,7 @@ SkyPilot can scale up the service to multiple service replicas with built-in aut
ports: 8081 # Expose to internet traffic.
envs:
PYTHONUNBUFFERED: 1
MODEL_NAME: meta-llama/Meta-Llama-3-8B-Instruct
HF_TOKEN: <your-huggingface-token> # Change to your own huggingface token, or use --env to pass.
@ -146,9 +147,8 @@ SkyPilot can scale up the service to multiple service replicas with built-in aut
run: |
conda activate vllm
echo 'Starting vllm api server...'
python -u -m vllm.entrypoints.openai.api_server \
vllm serve $MODEL_NAME \
--port 8081 \
--model $MODEL_NAME \
--trust-remote-code \
--tensor-parallel-size $SKYPILOT_NUM_GPUS_PER_NODE \
2>&1 | tee api_server.log
@ -243,6 +243,7 @@ This will scale the service up to when the QPS exceeds 2 for each replica.
ports: 8081 # Expose to internet traffic.
envs:
PYTHONUNBUFFERED: 1
MODEL_NAME: meta-llama/Meta-Llama-3-8B-Instruct
HF_TOKEN: <your-huggingface-token> # Change to your own huggingface token, or use --env to pass.
@ -258,9 +259,8 @@ This will scale the service up to when the QPS exceeds 2 for each replica.
run: |
conda activate vllm
echo 'Starting vllm api server...'
python -u -m vllm.entrypoints.openai.api_server \
vllm serve $MODEL_NAME \
--port 8081 \
--model $MODEL_NAME \
--trust-remote-code \
--tensor-parallel-size $SKYPILOT_NUM_GPUS_PER_NODE \
2>&1 | tee api_server.log

View File

@ -69,6 +69,11 @@ Sometimes you may see the API server entrypoint used directly instead of via the
python -m vllm.entrypoints.openai.api_server --model <model>
```
!!! warning
`python -m vllm.entrypoints.openai.api_server` is deprecated
and may become unsupported in a future release.
That code can be found in <gh-file:vllm/entrypoints/openai/api_server.py>.
More details on the API server can be found in the [OpenAI-Compatible Server](../serving/openai_compatible_server.md) document.

View File

@ -242,30 +242,8 @@ Example: `python3 -m tests.kernels.moe.modular_kernel_tools.profile_modular_kern
## FusedMoEPrepareAndFinalize Implementations
The following table lists the `FusedMoEPrepareAndFinalize` implementations at the time of writing,
| Implementation | Type | Comments |
| :--- | :--- | :--- |
| DeepEPHTPrepareAndFinalize | Contiguous / Non-Batched | Uses the DeepEP High-Throughput all2all kernels. |
| DeepEPLLPrepareAndFinalize | Batched | Uses the DeepEP Low-Latency all2all kernels. |
| PplxPrepareAndFinalize | Batched | Uses the Perplexity all2all kernels. |
| FlashInferCutlassMoEPrepareAndFinalize | Contiguous | |
| MoEPrepareAndFinalizeNoEP | Contiguous | This implementation is used when there is no EP. i.e. no all2all kernels are invoked. |
| BatchedPrepareAndFinalize | Batched | A reference prepare/finalize class that reorganizes the tokens into expert batched format, i.e. E x max_num_tokens x K. (Doesnt use any all2all kernels. This is primarily used in unit testing) |
See [Fused MoE Kernel features](./moe_kernel_features.md#fused-moe-modular-all2all-backends) for a list of all the available modular prepare and finalize subclasses.
## FusedMoEPermuteExpertsUnpermute
The following table lists the `FusedMoEPermuteExpertsUnpermute` implementations at the time of writing,
| Implementation | Type | Comment |
| :--- | :--- | :--- |
| BatchedDeepGemmExperts | Batched | Uses the DeepGemms Masked Grouped Gemm kernels for the fused_moe operation. |
| BatchedTritonExperts | Batched | Uses a Triton Kernel for the Batched matmuls. |
| BatchedTritonOrDeepGemmExperts | Batched | Chooses either the `BatchedDeepGemmExperts` or `BatchedTritonExperts` based on environment settings. |
| DeepGemmExperts | Contiguous / Non-Batched | Uses DeepGemms Grouped Gemm kernels for fused_moe operation. |
| TritonExperts | Contiguous / Non-Batched | Uses a Triton Kernel for fused_moe matmuls. |
| TritonOrDeepGemmExperts | Contiguous / Non-Batched | Chooses either the `DeepGemmExperts` or `TritonExperts` based on fused_moe inputs. |
| CutlassExpertsFP8 | Supports both Batched and Contiguous formats | Uses Cutlass Grouped Gemm implementations for the fp8 matmuls. |
| CutlassExpertsFP4 | Supports both Batched and Contiguous formats | Uses Cutlass Grouped Gemm implementations for the fp4 matmuls. |
| FlashInferExperts | Contiguous | Uses fused_moe operation from FlashInfer |
| NaiveBatchedExperts | Batched | Reference Batched Experts implementation. Primarily used in unit tests. |
See [Fused MoE Kernel features](./moe_kernel_features.md#fused-moe-experts-kernels) for a list of all the available modular experts.

Some files were not shown because too many files have changed in this diff Show More