Cheng Wan
1d47952c7c
config: pass the Ray placement group as a launch argument
...
`RayEngine.__init__` set the caller's `PlacementGroup` on `ServerArgs` as an
undeclared attribute, and every consumer read it back off the config object:
the non-DP launch path took `pg` and `is_custom_pg` from it, the DP path had to
re-attach the handle after `dataclasses.replace` dropped it, the DP controller
asked the config whether the group was the caller's, and the Ray HTTP launcher
had to clear the field so a stale handle could not leak in.
A live cluster object is not config. It now travels as a `placement_group`
argument on the two launch hooks — `Engine._launch_subprocesses` and
`Engine._launch_scheduler_processes` — defaulting to `None`, with `RayEngine`
holding the caller's group in `_placement_group` and the DP path passing
`is_custom_pg` down to the controller. The public API is unchanged:
`RayEngine(placement_group=pg, ...)` still works, `pg` still falls back to the
ambient group, and the HTTP launcher gets `None` from the default.
Writer ratchet 18 -> 15.
2026-08-05 19:28:39 -07:00
Michael
d90ef69802
[AMD] ci: run vetted nested multimodal_gen unit tests on AMD ( #31483 )
2026-08-06 09:39:33 +08:00
Pavan Sivaram Girijala
d9b1cba255
Fix DeepSeek-OCR batching crash on variable local-crop counts ( #33214 )
2026-08-05 18:39:19 -07:00
Zilin Zhu
f01f706960
[RL] Skip rotary cache tensors in weight checker ( #27692 )
2026-08-05 18:22:12 -07:00
Eric Zhang
ceaeca0b9e
[Fix] Vocab out of bounds in DSpark for Inkling-Small ( #33748 )
2026-08-05 18:20:02 -07:00
DevashishLal-CB and Devashish Lal
fc74c35546
kernels: scalar scale A support for fp8_gemm ( #33469 )
...
Signed-off-by: Devashish Lal <devcode@fb.com >
Co-authored-by: Devashish Lal <devcode@fb.com >
2026-08-06 09:11:25 +08:00
Eric Zhang
269d51ed4b
[Fix] Inkling works with gs:// runai_streamer paths ( #33750 )
2026-08-05 18:09:17 -07:00
28848bfe7c
Fix Nightly NV CI ( #33564 )
...
Co-authored-by: Brayden Zhong <brayden@radixark.ai >
Co-authored-by: Baizhou Zhang <sobereddiezhang@gmail.com >
2026-08-05 18:06:47 -07:00
Khoa Pham and Claude Opus 5
beabc5949b
Enable MoE deferred finalize by default and drop its expert_weights dtype workaround ( #33618 )
...
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com >
2026-08-05 17:56:47 -07:00
Aditya Kamat and Brayden Zhong
c9506d023f
[Quantization] Route per-tensor FP8 checkpoints to FlashInfer on SM90 ( #33148 )
...
Co-authored-by: Brayden Zhong <b8zhong@uwaterloo.ca >
2026-08-06 00:54:10 +00:00
Xiaoyu Zhang and Claude Fable 5
4c0a8940fa
[Kernel] Unify BaseFusedOp and MultiPlatformOp dispatch ( #33205 )
...
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-08-06 08:52:09 +08:00
Xiaoyu Zhang and Mohammad Miadh Angkad
ba12a16a62
[diffusion] Prefer cuDNN SDPA over FA4 for dense attention on sm_100 (B200) ( #33655 )
...
Co-authored-by: Mohammad Miadh Angkad <176301910+mmangkad@users.noreply.github.com >
2026-08-06 08:49:10 +08:00
sglang-bot and sglang-bot
c952ee5ac1
chore: bump sgl-kernel version to 0.4.6 ( #33678 )
...
Co-authored-by: sglang-bot <sglang-bot@users.noreply.github.com >
2026-08-05 17:46:54 -07:00
Mohammad Miadh Angkad
65d5a0ec25
Support ModelOpt MXFP8 checkpoints ( #32538 )
2026-08-05 17:44:32 -07:00
Liangsheng Yin
3869fe556f
[CI] Collapse the EAGLE launch matrix and the scoring engine boots on the per-commit runners ( #33756 )
2026-08-05 16:43:19 -07:00
Mohammad Miadh Angkad
5424d2039c
[CI] Fix GLM-Image usage unit tests ( #33731 )
2026-08-05 16:20:08 -07:00
Polisetty V R K Jyothendra Varma
a1cc286062
[Intel GPU] DeepSeek V4 4/N: use sgl-kernel implementation of fused_q_norm_rope on XPU ( #27790 )
...
Signed-off-by: P V R K Jyothendra Varma <polisettyvarma@gmail.com >
2026-08-05 16:04:35 -07:00
Liangsheng Yin
c0ef548eef
[misc] Unify MLA scaling init and remove dead buffer / scaling code ( #33363 )
2026-08-05 15:59:33 -07:00
YAMY
990a446773
Fix padded positions in breakable CUDA Graph attention ( #33253 )
...
TestBreakableCUDAGraphBasic and all NVIDIA CI tests pass.
2026-08-05 15:53:28 -07:00
YAMY and Chunan Zeng
7bc90ab394
Select DeepGEMM standard layouts by memory budget ( #33474 )
...
Co-authored-by: Chunan Zeng <zcnrex@gmail.com >
2026-08-05 15:50:43 -07:00
yvbbrjdr and hnyls2002
02cd44c59a
feat(dgx-spark): add inkling-small MoE support for sm_121 ( #33108 )
...
Co-authored-by: hnyls2002 <lsyincs@gmail.com >
2026-08-05 15:06:59 -07:00
cctry
a3a1ebc7b7
Warn on risky serving-time Triton work ( #33120 )
2026-08-05 14:55:03 -07:00
Ziang Li
988c6e6aeb
Pin online NVFP4 4over6 quantization settings ( #33621 )
2026-08-05 21:29:32 +00:00
Brayden Zhong and Brayden Zhong
a14c870886
Fix broken Nemotron DP attention ( #33123 )
...
Co-authored-by: Brayden Zhong <brayden@radixark.ai >
2026-08-05 14:27:05 -07:00
Hank Han
9436de717f
[Spec][PD] Enable fused TopK for GLM-5.2 MTP IndexShare ( #31477 )
2026-08-05 14:17:42 -07:00
Zhiqiang Xie
106bcc1293
Observability enhancement for HiCache ( #32388 )
2026-08-05 14:13:06 -07:00
Shu Wang
55b1c09e73
[core] Consolidate compiled-kernel caches under SGLANG_CACHE_DIR ( #32434 )
2026-08-05 13:54:27 -07:00
Jialin Ouyang
717a559f02
[Scheduler] Align WAR fences with CUDA graph metadata reads ( #33587 )
2026-08-05 13:52:20 -07:00
ea65f8ddc9
Feat/spectrum ( #31491 )
...
Co-authored-by: jaron1990 <jaron1990@gmail.com >
Co-authored-by: jaron1990 <34618972+jaron1990@users.noreply.github.com >
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com >
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com >
Co-authored-by: alexnails <alex.nails@radixark.ai >
2026-08-05 13:23:17 -07:00
5c4f72f92a
[Build] Add srt_empty extra group for device-agnostic install ( #31300 )
...
Co-authored-by: yixiaodapeng <yixiaodapeng@example.com >
Co-authored-by: Alex Nails <alex.nails@radixark.ai >
2026-08-05 13:17:46 -07:00
Liangsheng Yin
de34dd11e9
[CI] Fold duplicate-server suites and prune the retract matrix on 1-gpu-5090 ( #33745 )
2026-08-05 12:41:51 -07:00
Jason Mancuso
36853b8ffc
[Spec] Support logprobs with DFlash ( #33459 )
2026-08-05 12:37:12 -07:00
Khoa Pham
5f79cf3511
[DCP] Match the replicated draft KV pool's page granularity to its allocator ( #33348 )
2026-08-05 11:40:28 -07:00
Jialin Ouyang
b1bd871df5
[Unified Radix Cache] Complete the tree-core interface boundary ( #33580 )
2026-08-05 11:39:28 -07:00
cctry
96c89863a3
Measure prefill busy time between launches ( #33595 )
2026-08-05 11:26:27 -07:00
WenhaoZhang
acaab22d09
[diffusion] feat: add SageAttention packed varlen path for minimax-h3 ( #33703 )
2026-08-06 01:19:58 +08:00
zhaozx-cn
4e7209caa8
[NPU] Add causal conv1d ( #28267 )
2026-08-05 22:22:49 +08:00
Xiaoyu Zhang and Claude Fable 5
3425c93666
[diffusion] Wan VAE RMSNorm+SiLU fusion behind quality=high (H200 FastWan2.2 e2e 9.611 -> 9.125 s) ( #33546 )
...
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-08-05 21:33:35 +08:00
silencejade
593777c046
[FIX] [benchmark] Fix flush_cache failure after warmup by waiting for server idle ( #33527 )
2026-08-05 21:27:43 +08:00
Xuan Liao and Ma Mingfei
3b4fac5b99
[XPU] DeepSeek V4: use sgl-kernel-xpu implemetation of flash_mla_sparse_fwd for prefill ( #31865 )
...
Co-authored-by: Ma Mingfei <mingfei.ma@intel.com >
2026-08-05 21:05:53 +08:00
Mick
99709f734d
[VLM] split multimodal scheduling from mm_utils ( #32415 )
2026-08-05 20:24:12 +08:00
Xiaoyu Zhang and Claude Fable 5
a5888c956f
[diffusion] Pack Ulysses Q/K/V input all-to-all into one collective + reusable a2a staging buffers ( #33667 )
...
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-08-05 19:15:46 +08:00
2f22ed58ea
[NPU] Adding a fast layernorm for diffusion models and fix BSA ( #29027 )
...
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: ronnie_zheng <zl19940307@163.com >
2026-08-05 14:06:00 +03:00
22d558b103
[Feature] Add GLM Image usage report ( #33378 )
...
Co-authored-by: wuyuefeng <wuyuefeng@noreply.gitcode.com >
Co-authored-by: ronnie_zheng <zl19940307@163.com >
2026-08-05 13:59:49 +03:00
Alex Nails and Claude Fable 5
6fa3f9df11
[Bugfix] Treat unsharded model.safetensors as HF weights in Mistral-native format detection ( #33671 )
...
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
2026-08-05 01:54:46 -07:00
a6e5fa7081
[Scheduler] Honor explicit min-free-slots thresholds ( #33403 )
...
Co-authored-by: hnyls2002 <lsyincs@gmail.com >
Co-authored-by: Liangsheng Yin <hnyls2002@gmail.com >
2026-08-05 01:44:18 -07:00
Trevor Morris
81c7a54ecd
[NVIDIA] Use sm_100f instead of sm_100a for sgl-kernel and FlashMLA ( #33433 )
2026-08-05 01:36:46 -07:00
Xinyi Song
1478cdec9f
[AMD] Fuse Kimi-K3 attn-residual aggregation ( #33599 )
...
HIP Gated changes
2026-08-04 23:20:04 -07:00
Артем Савкин and Xiaoyu Zhang
d96df7bed5
[Diffusion] Batch GLM-Image AR requests ( #30683 )
...
Co-authored-by: Xiaoyu Zhang <1182563586@qq.com >
2026-08-05 08:47:06 +03:00
059269594c
[DSV4] Add official DSV4 reasoning effort support ( #33140 )
...
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com >
Co-authored-by: David Orman <ormandj@corenode.com >
2026-08-05 12:50:41 +08:00
Cheng Wan