Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Apr 23, 2026

• [None][feat] Move DWDP contention optimization into DwdpConfig (#12974)

  Show description for 79396d8
  JintaoPengCS

  authored
  79396d8

  Copy full SHA for 79396d8

• [https://nvbugs/6078421][fix] Commit dcb4a71b3 intentionally disabled `initialize_mrope_delta_cache` in `qwen3 (#13269)

  Show description for 56b0cbf
  tensorrt-cicd

  authored
  56b0cbf

  Copy full SHA for 56b0cbf

• [TRTLLMINF-45][infra] Upload CI agent failure analysis to SwiftStack (#13291)

  Show description for 43cd23f
  dpitman-nvda

  authored
  43cd23f

  Copy full SHA for 43cd23f

• [None][perf] Clear multimodal data upon prefill completion (#13259)

  Show description for 837a097
  2ez4bz

  authored
  837a097

  Copy full SHA for 837a097

• [#13125][feat] Make auto_deploy standalone-ready and add package generator (#13155)

  Show description for a492e5b
  lucaslie

  authored
  a492e5b

  Copy full SHA for a492e5b

• [None][fix] initialize sampler state for ADP dummy requests (#13275)

  Show description for 27f3382
  bobboli

  authored
  27f3382

  Copy full SHA for 27f3382

• [https://nvbugs/5919796][fix] AutoDeploy: Fix TP deadlock in multistream MoE (#13220)

  Show description for e1d4cf6
  galagam

  authored
  e1d4cf6

  Copy full SHA for e1d4cf6

• [None][fix] Disable multi stream moe for super v3 (#13122)

  Show description for a5244ae
  tcherckez-nvidia

  authored
  a5244ae

  Copy full SHA for a5244ae

• [None][doc] Restructure installation documentation (#12402)

  Show description for bf3b51d
  bobboli

  authored
  bf3b51d

  Copy full SHA for bf3b51d

• [https://nvbugs/6074943][fix] Disable new aiperf server metrics to stop hang. (#13314)

  Show description for b3a3b94
  dominicshanshan

  authored
  b3a3b94

  Copy full SHA for b3a3b94

• [None][chore] Add related trtllm-gen attention kernel files to trigger multi-gpu tests (#13260)

  heyuhhh

  authored
  f9f147c

  Copy full SHA for f9f147c

• [TRTLLM-11540][feat] Revert revert of EAGLE3 dynamic tree speculative decoding support (#13081)

  Show description for 3cc066b
  sunnyqgg

  authored
  3cc066b

  Copy full SHA for 3cc066b

• [TRTLLM-11485][feat] Feature rework: Add SageAttention refreshed kernels (attentionOp only) (#12937)

  Show description for 54c3915
  xrq-phys

  authored
  54c3915

  Copy full SHA for 54c3915

• [None][infra] Waive 2 failed cases for main in post-merge (#13335)

  Show description for cf20457
  xinhe-nv

  authored
  cf20457

  Copy full SHA for cf20457

• [https://nvbugs/6076560][chore] Unwaive test_nvfp4_4gpus (#13079)

  Show description for e3ca723
  hyukn

  authored
  e3ca723

  Copy full SHA for e3ca723

• [TRTLLM-12183][chore] Move bf16/auto precision variants from pre-merge to post-merge (#13366)

  Show description for 95d707f
  QiJune

  authored
  95d707f

  Copy full SHA for 95d707f

• [None][fix] Propagate init_load_balancer to DeepGemmFusedMoE in create_moe_backend (#13207)

  Show description for 5549b31
  qiaoxj07

  authored
  5549b31

  Copy full SHA for 5549b31

• [None][test] add models for weekly perf test (#13325)

  Show description for 7b0f60e
  ruodil

  authored
  7b0f60e

  Copy full SHA for 7b0f60e

• [None][chore] Better Empty File Error Handling for trtllm-bench (#12552)

  Show description for 36113bf
  

  yijingl-nvidia

  and

  claude

  authored
  36113bf

  Copy full SHA for 36113bf

• [None][infra] Check in most recent lock file from nightly pipeline

  Show description for 9b86622
  tensorrt-cicd

  committed
  9b86622

  Copy full SHA for 9b86622

• [https://nvbugs/6055847][fix] Preserve Nemotron HF mamba cache dtype in bench tuning (#12826)

  Show description for 23ff7d5
  hyukn

  authored
  23ff7d5

  Copy full SHA for 23ff7d5

• [https://nvbugs/6052050][fix] Drop stale CUDA graphs on beam-width change (#13255)

  Show description for 1d92d18
  brb-nv

  authored
  1d92d18

  Copy full SHA for 1d92d18

Commits on Apr 22, 2026

• [None][chore] Waive DSV32 tests (#13352)

  Show description for 659d5fb
  brb-nv

  authored
  659d5fb

  Copy full SHA for 659d5fb

• [https://nvbugs/6050481][chore] Unwaive passing GPT-OSS ep tests (#13284)

  Show description for 7a8bd87
  dongfengy

  authored
  7a8bd87

  Copy full SHA for 7a8bd87

• [None][chore] Add Dynamo configs to TRTLLM CI - Disagg - Part 2 (#13168)

  Show description for 519a7c7
  

  brb-nv

  and

  ZhanruiSunCh

  authored
  519a7c7

  Copy full SHA for 519a7c7

• [TRTLLM-11733][perf] Cache constant text computations across denoise steps in LTX2 (#12677)

  Show description for 29c8806
  luyiyun1021

  authored
  29c8806

  Copy full SHA for 29c8806

• [None][feat] Optimize causal_conv1d prefill and decode kernels (#13103)

  Show description for a8a8fa1
  Wanli-Jiang

  authored
  a8a8fa1

  Copy full SHA for a8a8fa1

• [TRTLLM-12137][chore] Drop non-key-model (starcoder2/mllama/nemotron) cases from L0 (#13315)

  Show description for a1bcae6
  QiJune

  authored
  a1bcae6

  Copy full SHA for a1bcae6

• [TRTLLM-12062][test] remove obsolete model tests (#13262)

  Show description for 62ce575
  xinhe-nv

  authored
  62ce575

  Copy full SHA for 62ce575

• [TRTLLM-11861][infra] Support wildcard in bot stage-list/extra-stage commands (#12881)

  Show description for 36fb5f0
  mzweilz

  authored
  36fb5f0

  Copy full SHA for 36fb5f0

• [None][fix] Enforce NCCL >= 2.28 at CMake configure time (#13108)

  Show description for 342adb3
  eopXD

  authored
  342adb3

  Copy full SHA for 342adb3

• [None][fix] Fix errors in KV cache manager V2 and scheduler V2 (#13104)

  Show description for 0d2bea7
  jiaganc

  authored
  0d2bea7

  Copy full SHA for 0d2bea7

• [None][infra] Check in most recent lock file from nightly pipeline

  Show description for 55419a9
  tensorrt-cicd

  committed
  55419a9

  Copy full SHA for 55419a9

• [None][fix] Fix post-merge perf data silently failing to upload to OpenSearch DB (#13215)

  Show description for f689e60
  chenfeiz0326

  authored
  f689e60

  Copy full SHA for f689e60

Commits on Apr 21, 2026

• [TRTLLM-11759][fix] Reduce peak host memory during NemotronH_Nano_VL_V2 init (#13283)

  Show description for a03eea6
  pamelap-nvidia

  authored
  a03eea6

  Copy full SHA for a03eea6