 Cheng Wan
|
e5a1c5a423
|
Fix _is_compiling dynamo tracing: import torch instead of sys.modules lookup (#36573)
|
2026-08-26 16:17:02 -07:00 |
|
|
kuma_Gu2006
|
7f27bf4708
|
[Kernel] Declare the PyTorch ABI dependency in sglang-kernel wheels (#36465)
|
2026-08-26 15:24:26 -07:00 |
|
|
Trevor Morris
|
1eb629ab4f
|
[NVIDIA] Tune custom all reduce v2 for sm_107 (#36397)
|
2026-08-26 15:19:31 -07:00 |
|
|
Trevor Morris
|
3ce243da3f
|
[NVIDIA] Add CUDA 13.4 container for initial Rubin support (#36233)
|
2026-08-26 15:02:16 -07:00 |
|
|
paulzhang-tm
|
06694071c6
|
[Spec] Avoid tensor scalar reads in spec decode allocation (#35377)
|
2026-08-26 14:51:31 -07:00 |
|
|
Xinyuan Tong
|
a8d716ce8d
|
dsa: widen the fp8 k-cache quant kernel's token_id to int64 (#30859)
|
2026-08-26 14:22:24 -07:00 |
|
|
Mohammad Miadh AngkadandMohammad Angkad
|
2935bb8e79
|
Fix OOB read in mxfp4 MoE weight scales on Hopper (#36456)
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai>
|
2026-08-26 13:36:49 -07:00 |
|
|
siyu
|
e7e7894016
|
[Weight Cache] Make daemon socket/ready paths configurable via env (#36299)
|
2026-08-26 12:38:50 -07:00 |
|
|
weireweireandweireweire
|
5263568bcb
|
[HiCache] Keep auxiliary load-back out of Full KV pending ownership (#36317)
Co-authored-by: weireweire <20922698+weireweire@users.noreply.github.com>
|
2026-08-27 01:30:38 +08:00 |
|
|
XuFuandMick
|
924aeee59c
|
[diffusion] feat: support batching for cosmos3 action generation (#36301)
Signed-off-by: FxxxxU <fu18801374388@163.com>
Signed-off-by: Mick <mickjagger19@icloud.com>
Co-authored-by: Mick <mickjagger19@icloud.com>
|
2026-08-26 22:15:03 +08:00 |
|
|
Cheng Wan
|
413df1f8db
|
config: ServerArgs holds the raw input (#36255)
|
2026-08-26 05:14:05 -07:00 |
|
|
|
27c36368b6
|
fix(moe): guard FP8 delegate activation params (#36275)
Signed-off-by: jikuixie <jikuixie@gmail.com>
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai>
Co-authored-by: Mohammad Miadh Angkad <176301910+mmangkad@users.noreply.github.com>
Co-authored-by: shyeh25 <206795756+shyeh25@users.noreply.github.com>
|
2026-08-26 12:12:21 +00:00 |
|
|
Cheng Wan
|
937af8538b
|
config: the runtime readers take the published bags (#36254)
|
2026-08-26 05:08:25 -07:00 |
|
|
Cheng Wan
|
5b7fc61306
|
config: resolution reads the declarations, not the fields (#36253)
|
2026-08-26 05:05:28 -07:00 |
|
|
Cheng Wan
|
ae5feb4b9c
|
config: stop handing the record to code that does not read it (#36252)
|
2026-08-26 05:02:17 -07:00 |
|
|
Cheng Wan
|
d7b144f64e
|
config: publishing is the process entry's job (#36251)
|
2026-08-26 04:58:41 -07:00 |
|
|
Xiaoyu Zhang
|
702de26310
|
[diffusion] make benchmark caches seedable and cover missing native families (#36463)
|
2026-08-26 19:04:46 +08:00 |
|
|
|
170da72c13
|
[diffusion] perf: fuse tanh-GELU into the LongCat-Image DiT FFN up-proj (#36322)
Co-authored-by: 登辉 <yangdenghui.ydh@alibaba-inc.com>
Co-authored-by: Xiaoyu Zhang <1182563586@qq.com>
|
2026-08-26 19:04:06 +08:00 |
|
|
Cheng Wan
|
8005df61d3
|
config: spell the parallel config tier at the call site (#36250)
|
2026-08-26 03:00:28 -07:00 |
|
|
DarkSharpness
|
689ade69d1
|
[kernel] Split the custom all-reduce communicator into push/pull planes (#35735)
|
2026-08-26 17:40:28 +08:00 |
|
|
Aurick QiaoandYuwei An
|
58ecbba0bd
|
[Feature] Coordinate FullCG prefill across DP-attention ranks (#35640)
Co-authored-by: Yuwei An <ayw.sirius19@gmail.com>
|
2026-08-26 02:16:02 -07:00 |
|
|
SuperSongandseokwoosong
|
2511743bd7
|
[unified-memory] Stop eviction when shared allocation capacity is sufficient (#33091)
Co-authored-by: seokwoosong <seokwoosong@users.noreply.github.com>
|
2026-08-26 02:06:32 -07:00 |
|
|
Mohammad Miadh AngkadandMohammad Angkad
|
a3c4936438
|
Sync FlashInfer autotune tactic choice across TP ranks (#35343)
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai>
|
2026-08-26 16:30:47 +08:00 |
|
|
KnightYao
|
7ef49e8f7d
|
Rename Spark3 to Spark2.5 (#36416)
|
2026-08-26 00:04:46 -07:00 |
|
|
|
3ce4f957eb
|
[diffusion] fix: fix MiniMax-H3 dp_size>1 deadlock and cross-request audio determinism (#36398)
Co-authored-by: Kevin Mi <kevin.mi@radixark.ai>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
|
2026-08-26 14:54:08 +08:00 |
|
|
Ziang Li
|
3c9febc68b
|
[Spec][DSA] Add --speculative-dsa-topk-backend (#36313)
|
2026-08-25 23:35:03 -07:00 |
|
|
weireweireandweireweire
|
db15976fe9
|
Fix DSV4 DSpark sample-from-anchor initialization (#36419)
Co-authored-by: weireweire <20922698+weireweire@users.noreply.github.com>
|
2026-08-25 23:18:58 -07:00 |
|
|
Tan TrinhandKhoa Pham
|
04c1036bb3
|
[Kernel] Skip reserved writes in MLA KV cache (#36003)
Co-authored-by: Khoa Pham <khoa.pham@radixark.ai>
|
2026-08-25 22:42:23 -07:00 |
|
|
Yihao WangandClaude Opus 5
|
d7baad0116
|
[diffusion] Keep the Cosmos3 Super DiT resident on high-memory GPUs (#36375)
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
2026-08-25 21:10:39 -07:00 |
|
|
|
34de1fb47f
|
fix(test): stabilize nightly precision regression (#34668)
Co-authored-by: Alison Shao <54658187+alisonshao@users.noreply.github.com>
Co-authored-by: Alison Shao <a.shao@wustl.edu>
|
2026-08-25 20:04:52 -07:00 |
|
|
ming_wang
|
2e4773aadd
|
Fix NPUMHATokenToKVPool missing k_data_ptrs/v_data_ptrs (#35820)
|
2026-08-26 11:01:11 +08:00 |
|
|
Bingxu Chen
|
5bcc978f2a
|
[AMD][CI] Pass USE_PDL explicitly in the fused MoE gate (#36306)
|
2026-08-25 19:57:59 -07:00 |
|
|
Xiaoyu Zhang
|
fa3ac61661
|
[Diffusion] Bound reusable Ulysses A2A staging buffers across shapes (#36327)
|
2026-08-26 10:57:29 +08:00 |
|
|
HuangJi
|
cc3b61873f
|
[Diffusion][minimax-h3] Restrict MiniMax-H3 SubBlock sparsity to video queries (#35850)
|
2026-08-26 10:55:09 +08:00 |
|
|
Yuanle Liu
|
4ae30dc736
|
fix: detect cross-node multimodal transport by nnodes (#35646)
|
2026-08-26 10:31:48 +08:00 |
|
|
|
2d8484740d
|
Support deepseek v4 and kimi k3 on ssd (#35314)
Co-authored-by: 1BIN4 <1741738350@qq.com>
Co-authored-by: L-Ark <fliangae@connect.ust.hk>
Co-authored-by: Chikati <jxudn@connect.ust.hk>
Co-authored-by: mengzili <zilim@ust.hk>
|
2026-08-26 10:05:12 +08:00 |
|
|
Zhangheng
|
bec6248272
|
[Unified Cache]: add glm5.2 per commit ci (#36281)
|
2026-08-26 10:04:19 +08:00 |
|
|
Joe
|
4b2c182d3f
|
[diffusion] feat: support out-of-tree torch.compile backends (#36249)
|
2026-08-26 09:53:39 +08:00 |
|
|
Mick
|
223dfce917
|
fix: bound CUDA memory for fast image preprocessing (#36295)
|
2026-08-26 09:02:08 +08:00 |
|
|
Baizhou ZhangandRyan Stewart
|
41e7612dee
|
[Model] Support Nemotron 3.5 Lightning speculative decoding (#36186)
Co-authored-by: Ryan Stewart <rystewart@nvidia.com>
|
2026-08-25 16:43:58 -07:00 |
|
|
Baizhou Zhang
|
2d88c79b3e
|
[CI] Add Kimi-K3 MMMU-Pro accuracy coverage (#36284)
|
2026-08-25 16:33:46 -07:00 |
|
|
cctry
|
aa718f7343
|
Refactor HiCache host pool management (#36232)
|
2026-08-25 16:31:57 -07:00 |
|
|
|
0c42a44cd7
|
Add Spark3 Model (#35963)
Co-authored-by: Yaowj <yaowj@MacBook-Air.local>
Co-authored-by: ybyang <10629930+whybeyoung@users.noreply.github.com>
|
2026-08-25 16:24:00 -07:00 |
|
|
Leon Gao
|
f7a56494b1
|
Fix SWA ownership across grouped frees (#36381)
|
2026-08-25 14:57:30 -07:00 |
|
|
Junpan Wu
|
4b4bf3d2a5
|
[Deepseek-V4] Enable shared-experts fusion on the flashinfer_mxfp4 (trtllm-gen) MoE path (#35505)
Signed-off-by: Shiki Wu <shikiw@nvidia.com>
|
2026-08-25 14:55:25 -07:00 |
|
|
Shangming Cai
|
edff717ef0
|
fix(disagg): snapshot affected rooms before iterating outside the lock (#36351)
|
2026-08-26 03:31:10 +08:00 |
|
|
Kevin FlansburgandShangming Cai
|
7ddf92d5f4
|
fix(disagg): refresh stale prefill bootstrap metadata (#36029)
Co-authored-by: Shangming Cai <csmthu@gmail.com>
|
2026-08-26 03:29:20 +08:00 |
|
|
YAMY
|
e9c9df6a52
|
[Performance] Tune FlashInfer EXTEND for DP prefill (#36219)
|
2026-08-25 08:29:57 -07:00 |
|
|
elvischenv
|
46d9427b91
|
Fix MXFP8 MoE weight sizing for non-gated models (#36097)
|
2026-08-25 22:09:09 +08:00 |
|
|
Benjamin Truong
|
a1f9508dd4
|
[Bugfix] Fix int32 destination offset overflow in CUTLASS MoE pre-reorder (#35188)
|
2026-08-25 20:56:42 +08:00 |
|