kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-02-05 02:31:33 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity
4,850 Commits 28 Branches 65 Tags 2 GiB
d6f76d2fae
Commit Graph

87 Commits

Author SHA1 Message Date
Chuang Zhu
d6f76d2fae
[TRTLLM-9527][feat] change context params and disagg params (step3) (#10495) 
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
 2026-01-27 16:34:17 +08:00
Linda
ce556290c9
[None][chore] Removing pybind11 bindings and references (#10550) 
Signed-off-by: Linda-Stadter <57756729+Linda-Stadter@users.noreply.github.com>
 2026-01-26 08:19:12 -05:00
Yao Yao
6f07fa81d7
[TRTLLM-7738][feat] Adding implementation of KVCacheManagerV2 (#10736) 
Signed-off-by: Yao Yao <lowsfer@users.noreply.github.com>

KVCacheManagerV2 is a new python-based implementation of the KV cache manager, featuring cleaner API, better abstraction and better code quality without the accumulated legacy.
 2026-01-24 04:48:39 -05:00
Leslie Fang
31d04dfa12
[TRTLLM-9108][feat] Add test configurable moe module multi gpu (#10699) 
Signed-off-by: leslie-fang25 <leslief@nvidia.com>
 2026-01-23 10:16:58 +08:00
Yi Zhang
d43be7b65e
[None][fix] Avoid Double update for previous batch (#9888) 
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
 2026-01-22 13:15:06 -05:00
Jiayu Chang
1dc49b266e
[https://nvbugs/5322131][feat] Multi-LoRA serving with CUDA Graph (#8279) 
Signed-off-by: Jiayu Chang <jiayuc@nvidia.com>
 2026-01-22 14:01:18 +01:00
Lizhi Zhou
f3a41c8d94
[TRTLLM-10059][feat] Use global unique id as disagg request id (#10187) 
Signed-off-by: Lizhi Zhou <1432185+reasonsolo@users.noreply.github.com>
 2026-01-21 22:52:34 -05:00
jthomson04
2db3d7eeba
[None][chore] Async Transfer Manager (#9891) 
Signed-off-by: jthomson04 <jwillthomson19@gmail.com>
 2026-01-20 12:12:47 -05:00
Yi Zhang
58311b2345
[None][fix] Remove unused params in attn (#10652) 
Signed-off-by: yizhang-nv <187001205+yizhang-nv@users.noreply.github.com>
 2026-01-20 03:08:59 -05:00
Liao Lanyu
dbb858ae0c
[TRTLLM-10029][scheduler] Re-implement MicroBatchScheduler and CapacityScheduler in Python (#10273) 
Signed-off-by: junq <22017000+QiJune@users.noreply.github.com>
Signed-off-by: Lanyu Liao <lancelly@users.noreply.github.com>
Signed-off-by: Lance Liao <108499334+lancelly@users.noreply.github.com>
Co-authored-by: junq <22017000+QiJune@users.noreply.github.com>
Co-authored-by: Lanyu Liao <lancelly@users.noreply.github.com>
 2026-01-20 10:31:13 +08:00
Yukun He
f001c4946d
[https://nvbugs/5782112][fix] Fix hanging issue for MNNVL Allreduce under PP (#10633) 
Signed-off-by: Yukun He <23156053+hyukn@users.noreply.github.com>
 2026-01-16 13:03:36 +08:00
Lizhi Zhou
93db0d5e18
[TRTLLM-9942][feat] new request states and kvcache transceiver APIs in generation-first disagg (#10406) 
Signed-off-by: Lizhi Zhou <1432185+reasonsolo@users.noreply.github.com>
 2026-01-15 19:18:21 +08:00
Iman Tabrizian
48b09e5a25
[https://nvbugs/5689235][fix] Fix cancellation+chunked prefill+disagg (#10111) 
Signed-off-by: Iman Tabrizian <10105175+tabrizian@users.noreply.github.com>
 2026-01-12 18:23:26 -05:00
Chuang Zhu
536a8f6a9c
[TRTLLM-9527][feat] Add transferAgent binding (step 1) (#10113) 
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
 2026-01-06 08:40:38 +08:00
Jin Li
c04563657e
[TRTLLM-7735][feat] Attention NVFP4 out support for torch compile (#9740) 
Signed-off-by: Jin Li <59594262+liji-nv@users.noreply.github.com>
 2025-12-27 00:07:20 +08:00
Balaram Buddharaju
8c1cfc872b
[TRTLLM-9493][feat] Custom AllToAll for helix parallelism (#9986) 
Signed-off-by: Balaram Buddharaju <169953907+brb-nv@users.noreply.github.com>
 2025-12-23 18:14:30 -08:00
Bo Li
a66eeab537
[TRTLLM-9805][feat] Skip Softmax Attention. (#9821) 
Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com>
Signed-off-by: Tian Zheng <29906817+Tom-Zheng@users.noreply.github.com>
Co-authored-by: Tian Zheng <29906817+Tom-Zheng@users.noreply.github.com>
 2025-12-21 02:52:42 -05:00
Enwei Zhu
2ce785f39a
[https://nvbugs/5643631][fix] Fix hostfunc seg fault (#10028) 
Signed-off-by: Enwei Zhu <21126786+syuoni@users.noreply.github.com>
 2025-12-20 07:58:43 -05:00
Wangjue Yao
9f283f330b
[None][feat] Support Mooncake transfer engine as a cache transceiver backend (#8309) 
Signed-off-by: wjueyao <wyao123@terpmail.umd.edu>
Signed-off-by: Shunkang <182541032+Shunkangz@users.noreply.github.co>
Co-authored-by: Shunkang <182541032+Shunkangz@users.noreply.github.co>
 2025-12-19 10:09:51 +08:00
Simeng Liu
f21e2b3329
[TRTLLM-9601][feat] Expose mmKeys for multimodal to integrate with dynamo. (#9604) 
Signed-off-by: SimengLiu-nv <simengl@nvidia.com>
 2025-12-15 08:42:30 +08:00
Yihan Wang
9df4dad3b6
[None][fix] Introduce inline namespace to avoid symbol collision (#9541) 
Signed-off-by: Yihan Wang <yihwang@nvidia.com>
 2025-12-12 23:32:15 +08:00
Iman Tabrizian
9425f7fe3a [https://nvbugs/5601682][fix] Fix cacheTransceiver hang (#9311) 
Signed-off-by: Iman Tabrizian <10105175+tabrizian@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com>
Signed-off-by: Mike Iovine <miovine@nvidia.com>
 2025-12-05 17:50:12 -05:00
Thor Johnsen
95049eea86
[https://nvbugs/5627710][fix] Fix synchronization bugs in KvCacheTransferManager that can cause corrupted blocks (#9056) 
Signed-off-by: thorjohnsen <41591019+thorjohnsen@users.noreply.github.com>
Signed-off-by: Thor Johnsen <41591019+thorjohnsen@users.noreply.github.com>
Co-authored-by: Iman Tabrizian <10105175+tabrizian@users.noreply.github.com>
Co-authored-by: Robin Kobus <19427718+Funatiq@users.noreply.github.com>
 2025-12-02 09:10:21 -06:00
Bo Li
fcfec93cad
[TRTLLM-9389][chore] Rename AlltoAll backend names (#9329) 
Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com>
 2025-11-23 13:52:57 -08:00
Patrice Castonguay
9b0f45298f
[None][feat] Have ability to cancel disagg request if KV cache resource are exhausted (#9155) 
Signed-off-by: Patrice Castonguay <55748270+pcastonguay@users.noreply.github.com>
 2025-11-18 20:59:17 -05:00
Robin Kobus
9913dc25ae
[None][refactor] decoding inputs, part 2 (#5799) 
Signed-off-by: Robin Kobus <19427718+Funatiq@users.noreply.github.com>
 2025-11-18 14:38:51 +01:00
Robin Kobus
df41f220a2
[TRTLLM-8831][feat] Enable early exit with overlap scheduler (#8587) 
Signed-off-by: Robin Kobus <19427718+Funatiq@users.noreply.github.com>
 2025-11-17 18:07:13 +01:00
heyuhhh
f07e9977c6
[None] [feat] Use triton kernels for RocketKV prediction module (#8682) 
Signed-off-by: yuhangh <58161490+heyuhhh@users.noreply.github.com>
 2025-11-13 18:51:09 -08:00
yunruis
51545560da
[TRTLLM-8803][feat] Add rope and uk-bgemm overlap for mla generation (#8495) 
Signed-off-by: yunruis <205571022+yunruis@users.noreply.github.com>
 2025-11-06 17:39:57 +08:00
shuyixiong
70e4d72ffa
[TRTLLM-8511][feat] Add update_weights and sleep_wakeup support for rl integration (#8302) 
Signed-off-by: shuyix <219646547+shuyixiong@users.noreply.github.com>
Co-authored-by: Liwei Ma <liweim@nvidia.com>
Co-authored-by: Jonas Yang CN <joyang@nvidia.com>
 2025-11-04 10:19:24 -08:00
Bo Li
e4bf29bc66
[None][feat] Integrate MnnvlThroughput into TRTLLM MoE. (#8728) 
Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com>
 2025-11-04 21:36:29 +08:00
Fanrong Li
f0dc746738
[TRTLLM-8541][feat] Add trtllm-gen sparse MLA kernels to support per-Tensor FP8 KV Cache (#8692) 
Signed-off-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>
Signed-off-by: Tracin <10434017+Tracin@users.noreply.github.com>
Signed-off-by: Fanrong Li <23290157+lfr-0531@users.noreply.github.com>
Co-authored-by: Perkz Zheng <67892460+PerkzZheng@users.noreply.github.com>
Co-authored-by: Tracin <10434017+Tracin@users.noreply.github.com>
 2025-10-31 14:38:31 -07:00
Iman Tabrizian
ae6875fe10
[TRTLLM-8976][feat] Move indexer-k-cache to KVCacheManager (#8699) 
Signed-off-by: Iman Tabrizian <10105175+tabrizian@users.noreply.github.com>
 2025-10-29 08:04:26 -07:00
Zheng Duan
fea5bfbda7
[None][feat] add detailed KV cache transfer time breakdown (#8521) 
Signed-off-by: zhengd-nv <200704041+zhengd-nv@users.noreply.github.com>
 2025-10-29 10:11:09 +08:00
Bo Li
9c4432f8a4
[TRTLLM-7318][feat] MnnvlThroughput AlltoAll implementation. (#7499) 
Signed-off-by: Bo Li <22713281+bobboli@users.noreply.github.com>
Co-authored-by: Jin Li <59594262+liji-nv@users.noreply.github.com>
 2025-10-27 13:23:06 -04:00
nvxuanyuc
d1398c05e6
[None][feat] Support ignored prompt length for penalties via new sampling config parameter (#8127) 
Signed-off-by: Xuanyu Chen <xuanyuc@nvidia.com>
 2025-10-27 13:12:31 -04:00
Chuang Zhu
2420918e5b
[TRTLLM-7078][chore] optimal kvcache transfer for VWSA (#7952) 
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
 2025-10-24 08:58:16 -04:00
Patrice Castonguay
879039f6d5
[https://nvbugs/5429636][feat] Kv transfer timeout (#8459) 
Signed-off-by: raayandhar <raayan.dhar@gmail.com>
Signed-off-by: Patrice Castonguay <55748270+pcastonguay@users.noreply.github.com>
Co-authored-by: raayandhar <raayan.dhar@gmail.com>
 2025-10-22 09:29:02 -04:00
Chuang Zhu
40d129a415
[None][fix] Fix cache buffer size for window (#8320) 
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
 2025-10-16 09:01:11 +08:00
Fanrong Li
0d20a8fd61
[TRTLLM-8536][feat] Add the sparse attention framework and one use case--RocketKV support (#8086) 
Signed-off-by: Fanrong Li <23290157+lfr-0531@users.noreply.github.com>
Signed-off-by: yuhangh <58161490+heyuhhh@users.noreply.github.com>
Co-authored-by: yuhangh <58161490+heyuhhh@users.noreply.github.com>
 2025-10-14 08:23:16 -07:00
Jonas Yang CN
88ea2c4ee9
[TRTLLM-7349][feat] Adding new orchestrator type -- ray (#7520) 
Signed-off-by: Erin Ho <14718778+hchings@users.noreply.github.com>
Co-authored-by: Yuan Tong <13075180+tongyuantongyu@users.noreply.github.com>
Co-authored-by: Erin Ho <14718778+hchings@users.noreply.github.com>
 2025-10-04 08:12:24 +08:00
Yilin Fan
01423ac183
[None][feat] perf_metrics endpoint functionality improvement (#8005) 
Signed-off-by: Yilin Fan <206948969+nv-yilinf@users.noreply.github.com>
Signed-off-by: nv-yilinf <206948969+nv-yilinf@users.noreply.github.com>
 2025-10-02 17:43:25 -07:00
Patrice Castonguay
fefa7d8fa3
[None][feat] Support for cancelling requests with disaggregation (#8114) 
Signed-off-by: Shunkang <182541032+Shunkangz@users.noreply.github.co>
Signed-off-by: Patrice Castonguay <55748270+pcastonguay@users.noreply.github.com>
Co-authored-by: Shunkang <182541032+Shunkangz@users.noreply.github.co>
 2025-10-02 11:04:26 -07:00
Iman Tabrizian
33282351a2
[TRTLLM-6106][feat] Add support for KVCache transfer from KVCache reuse path (#6348) 
Signed-off-by: Iman Tabrizian <10105175+tabrizian@users.noreply.github.com>
 2025-09-27 19:29:30 -04:00
Yuan Tong
fae83c387b
[#6102][fix] support non-system python installation (#7763) 
Signed-off-by: Yuan Tong <13075180+tongyuantongyu@users.noreply.github.com>
 2025-09-26 10:16:15 +08:00
Guoming Zhang
202bed4574 [None][chroe] Rename TensorRT-LLM to TensorRT LLM for source code. (#7851) 
Signed-off-by: nv-guomingz <137257613+nv-guomingz@users.noreply.github.com>
Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
 2025-09-25 21:02:35 +08:00
Yueh-Ting (eop) Chen
cf100933cc
[TRTLLM-6341][feature] Support SWA KV cache reuse (#6768) 
This merge request attempts to support more SWA KV cache functionality
inside the KV cache manager. Before this merge request, the KV cache for
sliding window attention (SWA) only holds "window size" number of blocks
and reuse them in a cyclic manner. We will not be able to utilize more
GPU memory with this design, leading to a limited max batch size
throughput. Additionally, we will not be able to support KV cache reuse
with this design.

In this MR, we change such behavior to let the manager write blocks in
a linear manner. With a linear block writing behavior, as the attention
window moves on, the out-of-window (OOW) blocks will be detached. Right
now for the sake of a correct feature first, we directly offload the
OOW block from the primary block pool (GPU memory) to the secondary
block pool (host memory). We will improve this in the future by
delegating the block movement to the eviction policy.

KV cache reuse for SWA is not developed in this merge request and will
be amended in a follow-up merge request.

Writing the blocks linearly, the maximum number of blocks allocated for
a sequence(`GenerationRequest`) is the "max sequence length" specified.
The `GenerationRequest` that stores the cache block bookkeeping
structure will now keep "max sequence length" tokens of blocks.

Given the above, main changes are (more context in the MR):
- Remove "cyclic" concept under the kv cache manager, such concept
  originally guards the block reuse under kv cache manager.
- Add detach mechanism and have it under `KVCacheManager::addToken`.
  Please note that detach is still guarded off for SWA when reuse
  is enabled. A follow-up merge request will proceed to improve this.
- Enforce "max sequence length" to be a non-optional parameter to
  the `KVCacheManager`/`BlockManager`
- Let all window size resource pool get identical proportion of memory
- Fix free memory calculation under `resource_manager.py`

Signed-off-by: eopXD <yuehtingc@nvidia.com>
Co-authored-by: Tomer Asida <tasida@nvidia.com>
 2025-09-24 14:28:24 +08:00
Matthias Jouanneaux
1be7faef37
[TRTLLM-5966][feat] Helix: add custom position ids to MLA kernels (#6904) 
Signed-off-by: Matthias Jouanneaux <mjoux@nvidia.com>
Co-authored-by: brb-nv <169953907+brb-nv@users.noreply.github.com>
 2025-09-19 20:55:32 +08:00
Shiyu Li
8bdbb48264
[https://nvbugs/5489015][fix] Support communicator split in MNNVL allreduce and fix the binding issues. (#7387) 
Signed-off-by: Shiyu Li <shili@nvidia.com>
 2025-09-17 07:43:20 +08:00
jmydurant
7deefb3d2b
[TRTLLM-7192][feat] optimize MLA chunked prefill && support fp8 mla chunked prefill (#7477) 
Signed-off-by: Mingyang Jiang <13463932+jmydurant@users.noreply.github.com>
 2025-09-15 21:43:49 +08:00