78 Commits
Author SHA1 Message Date
Jee Jee Li 39983a5e7a [Kernel] Add FP32 router GEMM support for Step-5 Preview and Muse Spark Open (#62126)
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-10-16 22:17:55 -07:00
596e530675 [Kernel] Add router GEMM kernels for (5120, 256) and BF16 output to ll_bf16_gemm (#62042)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: Lain <fusiyuan2000@gmail.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: vllm-agent <claw@inferact.ai>
2026-10-16 15:39:46 -07:00
Andreas KaratzasandClaude b466fbacd2 [ROCm][Perf] Any-shape low-latency MoE router GEMM on MI355 (#61895)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-10-15 23:22:31 -05:00
Mikko TukiainenandCursor 47fd76e4ce [ROCm][Perf] Route softmax MoE top-k, including fused shared-expert sigmoid, through AITER topk_gating (#56226)
Signed-off-by: Mikko Tukiainen <Mikko.Tukiainen@amd.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-10-15 17:53:18 +00:00
Chaunceyandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> ad8d39cbab [Bugfix] Fix DSpark graph-padding handling in the DeepSeek-V4.1 MoE fused top-k (#59166)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-10-14 07:49:52 +08:00
68fc395554 [Bugfix] Keep GateLinear batch invariant under VLLM_BATCH_INVARIANT (#59985)
Signed-off-by: yifanFengg <evanfengff@gmail.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
2026-10-13 15:43:54 -04:00
Roberto L. CastroandClaude Opus 5.5 7fad7db330 [Perf][Feat] Add generic cuteDSL LL FP32 router (GEMM) (#44343)
Signed-off-by: LopezCastroRoberto <rocastro@redhat.com>
Signed-off-by: Roberto L. Castro <38211239+LopezCastroRoberto@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-13 12:38:46 -04:00
akii96 182247257f [ROCm][Perf] Use a stacked hipBLASLt GEMM for the bf16x3 router (#52668)
Signed-off-by: Aakif Nawaz <aakif.nawaz@amd.com>
2026-10-05 17:28:07 +00:00
Sergey Shlyapnikov 4e1182a3a6 [ROCm][MoE] Pad the AITER MoE intermediate size at allocation time, and round the expert-group count to a kernel that exists (#55368)
Signed-off-by: Sergei Shliapnikov <sergei.shliapnikov@amd.com>
2026-09-30 23:08:21 +08:00
Jiangyun Zhu 231fdb83cc [Perf][MoE] Use fused MiniMax2 routing with non-unit routed scaling (#58880)
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
2026-09-27 19:37:33 +00:00
Nick Hill 09fe178dba Revert "[DSpark] Support pipeline-parallel targets in aggregated serving (#56956)" (#58484) 2026-09-24 08:48:36 -07:00
157bcb7c48 [DSpark] Support pipeline-parallel targets in aggregated serving (#56956)
Signed-off-by: Zihua Wu <13583761+lucifer1004@users.noreply.github.com>
Co-authored-by: Kimi Code <noreply@moonshot.cn>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>
2026-09-23 04:52:28 -07:00
Jee Jee Li de6674b002 [MoE] Use GateLinear for all MoE models (#58234)
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-09-23 05:40:13 +00:00
Elvir CrnčevićOpenAI Codexmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>Tyler Michael Smith
ac2f0ea82c [MoE][Bugfix] Skip SP padded rows in grouped MoE routing (#56079)
Signed-off-by: Elvir Crncevic <elvircrn@gmail.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2026-09-17 19:36:19 +00:00
Qiming Zhang e6b1a5e5e3 [XPU] Enable XPU eplb (#44987)
Signed-off-by: mayuyuace <qiming1.zhang@intel.com>
Signed-off-by: Qiming Zhang <qiming1.zhang@intel.com>
2026-09-17 13:19:13 +08:00
Harry MellorandClaude Opus 5 651a88c09a [CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 10:57:24 -07:00
Harry MellorandClaude Opus 5 8c1557a79c Add pydocstyle to the ruff rules (#52136)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Harry Mellor <hej.mellor@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-09-16 16:44:29 +00:00
9b959b8657 [Model] DeepSeek-V4.1-Flash Model Definitions (#56228)
Signed-off-by: Roger Wang <hey@rogerw.io>
Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Kimi Code <noreply@moonshot.cn>
Co-authored-by: zjy0516 <riverclouds.zhu@qq.com>
Co-authored-by: Codex <noreply@openai.com>
2026-09-10 20:16:26 +08:00
Li, Jiang c3ccc0e957 [Model] Add DeepSeek-V4 CPU backend (#55355)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
2026-09-10 14:14:48 +08:00
cc4210f671 [Perf] Improve BF16x3 router GEMM accuracy and make it default on sm100 (#55899)
Signed-off-by: Thien Tran <gau.nernst@yahoo.com.sg>
Co-authored-by: OpenAI Codex <noreply@openai.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
2026-09-09 07:33:12 -07:00
Marceli FylcekandClaude Opus 5 9b85112e13 [XPU] Route grouped_topk to the fused _moe_C kernel on XPU (#53580)
Signed-off-by: Marceli Fylcek <marceli.fylcek@intel.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-09-07 17:43:38 +08:00
Fangzhou Ai 8bf39632b8 [ROCm][Perf] Add low-M FP32 router GEMM for gfx950 (#54845) 2026-09-03 10:42:25 -07:00
Isotr0py 1356635d83 [New model][Multimodal] Add DeepSeek-V4-Flash-Vision-Exp support (#54566)
Signed-off-by: Isotr0py <Isotr0py@outlook.com>
2026-09-02 11:09:09 -07:00
Jee Jee Li 8e92248f79 [Kernel] Retire the DSv3 router GEMM CUDA kernel (#54040)
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-08-31 07:15:31 +00:00
hcl b5707bf994 [Bugfix][MoE] Enable cuBLAS out_dtype router GEMM on all CUDA archs (fixes family-120/GB10) (#54048)
Signed-off-by: Chenglun Hu <chenglunhu@gmail.com>
2026-08-29 19:18:42 -07:00
Matvei PashkovskiiandClaude 20727e2841 [Hardware][AMD] Enable fused bf16→fp32 router GEMM on ROCm (#50268)
Signed-off-by: Matvei Pashkovskii <mpashkov@amd.com>
Co-authored-by: Claude <noreply@anthropic.com>
2026-08-12 02:46:13 +00:00
Elvir CrnčevićClaude Opus 4.6gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>Tyler Michael Smith
8fea1d306f Fix uniform_random routing simulation to sample without replacement (#43680)
Signed-off-by: Elvir Crncevic <elvircrn@gmail.com>
Signed-off-by: Elvir Crnčević <elvircrn@gmail.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
2026-08-10 15:28:57 +00:00
Liangqiusong 1e05b21d61 Remove the XPU branch of topk_softplus_sqrt (#51242)
Signed-off-by: xiaolong <xiaolong.guo@intel.com>
2026-08-06 12:18:45 +00:00
Julien Debache bb3b61f2fd perf: dispatch non-grouped bias-less topk routing methods to fused path (#49618)
Signed-off-by: jdebache <jdebache@nvidia.com>
2026-07-28 14:57:22 -07:00
gnovack 85f638a2b8 skip cudagraph/DP padding in topk (#48979)
Signed-off-by: gnovack <novackgm@gmail.com>
2026-07-21 15:20:30 -07:00
Wentao Ye b5433b6f50 [Perf] Optimize dsv4 routing using specialized kernel, 2.94% E2E TPOT improvement (#48660)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2026-07-17 13:35:06 -07:00
Thien Tran 85e296950c BF16x3 router GEMM (#47973)
Signed-off-by: Thien Tran <gau.nernst@yahoo.com.sg>
2026-07-16 17:04:20 +08:00
Wentao Ye 2dab187f75 [Perf] Optimize fused_topk_bias for DSv4, 1.5~2x kernel performance improvement (#47463)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
2026-07-15 23:40:55 +00:00
Roberto L. CastroandLucas Wilkinson 0762f2afeb [Perf][Feat] Add generic cuteDSL LL BF16 router (GEMM) (#42562)
Signed-off-by: LopezCastroRoberto <rocastro@redhat.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Roberto L. Castro <38211239+LopezCastroRoberto@users.noreply.github.com>
Co-authored-by: Lucas Wilkinson <lwilkins@redhat.com>
2026-07-13 23:01:21 -07:00
19069bcbd5 FP32 router GEMV optimization (#48335)
Signed-off-by: peiyuanz <peiyuanz@inferact.ai>
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
Co-authored-by: peiyuanz <peiyuanz@inferact.ai>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: zhouzhou <zhouzhou@zhouzhoudeMacBook-Pro.local>
2026-07-11 13:07:48 +00:00
Jee Jee Li 5d23ca47ab [Kernel] Applies routed_scaling_factor internally (#47408)
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-07-07 02:00:54 -07:00
Yuxuan Zhang 178fd56094 support GLM-5.2 gate use FP32 (#47410)
Signed-off-by: zRzRzRzRzRzRzR <Yuxuan.Zhang2@liverpool.ac.uk>
2026-07-02 22:45:39 +08:00
Ilya MarkovandMarkov Ilya a2abce646f [EPLB] Mask padding in EPLB load recording (#38128)
Signed-off-by: ilmarkov <markovilya197@gmail.com>
Signed-off-by: Markov Ilya <markovilya19@gmail.com>
Co-authored-by: Markov Ilya <markovilya19@gmail.com>
2026-06-28 19:43:58 -07:00
Hongxia YangandClaude Opus 4.8 c2507fb293 [ROCm] [MoE] [Perf] Shared-expert fusion for bias-routed MoE; enable on MiniMax-M3 mxfp8 model (#46545)
Signed-off-by: Hongxia Yang <hongxia.yang@amd.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-26 07:05:20 -07:00
Jee Jee Li 9d6fdc2901 [Kernel] GLM5 Router GEMM (#46385)
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-06-23 22:54:50 -07:00
0a1c5034f5 [Model] Add MiniMax M3 support (#45381)
Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Bugen Zhao <i@bugenzhao.com>
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Signed-off-by: functionstackx <47992694+functionstackx@users.noreply.github.com>
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Thien Tran <gau.nernst@yahoo.com.sg>
Co-authored-by: Bugen Zhao <i@bugenzhao.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
Co-authored-by: functionstackx <47992694+functionstackx@users.noreply.github.com>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-06-16 01:01:25 +08:00
Roberto L. Castro 6850839c6f [Perf] Fix dsv3_router_gemm heuristic (#44217)
Signed-off-by: LopezCastroRoberto <rocastro@redhat.com>
2026-06-10 06:08:41 -07:00
Fangzhou AiandvLLM Contributor 01d8cd92dd [ROCm][Perf] Use fused softplus-sqrt-topk router under AITER fused-MoE (#44945)
Co-authored-by: vLLM Contributor <contributor@vllm.ai>
2026-06-09 17:53:05 +00:00
Tyler Michael SmithandClaude Opus 4.6 e2f993dc41 [WideEP] Integrate DeepEP v2 (#41183)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-06-08 18:07:29 -07:00
bnellnm dc68bd8c41 [MoE Refactor] FusedMoE/MoERunner inversion refactor (#41184)
Signed-off-by: Bill Nell <bnell@redhat.com>
2026-06-08 10:42:58 -04:00
Terrence Zhao a50e675b0d [Cohere] fix RoutingMethodType (#44021)
Signed-off-by: Terrencezzj <terrence@cohere.ai>
2026-06-05 16:25:53 -07:00
Wei Zhaoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> ace95c9cf8 [Bugfix] Update TrtLLM MoE routing methods (#44347)
Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-03 02:56:43 -07:00
Jee Jee LiandYiliu Dong 559d6710bf [PERF]MiniMax-M2 gate kernel (#38445)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Signed-off-by: qianlihuang <91178480+qianlihuang@users.noreply.github.com>
Co-authored-by: Yiliu Dong <91178480+qianlihuang@users.noreply.github.com>
2026-05-29 18:28:34 -07:00
Jee Jee Li d4004455d2 [Kernel] Remove NormGateLinear (#43554)
Signed-off-by: Jee Jee Li <jeejeelee@inferact.ai>
2026-05-25 09:49:19 +00:00
weizhoublue 6cbe448eed fix: MoE model using shared routed experts crashes on AMD GPUs (#42373)
Signed-off-by: weizhou.lan@daocloud.io <weizhou.lan@daocloud.io>
2026-05-25 12:03:05 +08:00