Pull requests: NVIDIA/TensorRT-LLM
Pull requests list
[#12502][fix] Export softmax stats with the real scale so SM90 MLA chunked prefill works
#19112
opened Sep 12, 2026 by
akhoroshev
Contributor
Loading…
[https://nvbugs/6746175][fix] Remove the now-obsolete
waives.txt SKIP for the GLM-5 tep8_mtp3 lane…
#19111
opened Sep 12, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
[None][fix] thread LoRA params through MoE models
#19109
opened Sep 12, 2026 by
achartier
Collaborator
Loading…
1 task done
[TRTLLM-16206][fix] Reduce KV fraction for Qwen3.5 OOM perf cases
#19108
opened Sep 12, 2026 by
VALLIS-NERIA
Collaborator
Loading…
1 task done
[None][test] Enable gen_only_no_context mode in perf sanity system
#19107
opened Sep 12, 2026 by
chenfeiz0326
Collaborator
Loading…
4 tasks done
[None][fix] KVCM2: restore Python/C++ parity checks and replace abort with a poison latch
#19103
opened Sep 12, 2026 by
lowsfer
Member
Loading…
1 task done
[nvbugs/6765038][fix] Pass required aux_stream_dict to KimiK3MoERuntime in SiTU NVFP4 test
#19102
opened Sep 12, 2026 by
HuiGao-NV
Collaborator
Loading…
1 task done
[https://nvbugs/6732110][fix] Cache the split communicator in a new `MNNVLAllReduce.allreduce_mnnvl_pending_c…
#19101
opened Sep 12, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
[None][infra] Waive 1 failed cases for main in pre-merge 59979
#19100
opened Sep 12, 2026 by
trtllm-agent
Collaborator
Loading…
[https://nvbugs/6731978][fix] Fix AllReduce autotuner ranking non-equivalent fused tactics on latency
#19099
opened Sep 12, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
[NVBUG-6762388][fix] enable Cosmos3 Edge Diffusers parity
VisualGen
#19097
opened Sep 12, 2026 by
ishovkun
Contributor
Loading…
[https://nvbugs/6641268][perf] Use producer-zeroed V tails in PrimTS FMHA
#19096
opened Sep 12, 2026 by
yuxianq
Collaborator
Loading…
1 task done
[None][fix] Keep the family suffix for 10x architectures in DeepEP
#19095
opened Sep 12, 2026 by
100milliongold
Loading…
3 of 4 tasks
[TRTLLM-14329][perf] Enable fused DSA metadata by default
#19094
opened Sep 12, 2026 by
xwang233
Collaborator
Loading…
[None][test] Fix tensor-sharing test handshake and xdist reporting
#19093
opened Sep 12, 2026 by
chang-l
Collaborator
Loading…
[https://nvbugs/6708299][fix] Support group_size=32 for W4A16_NVFP4 and validate NVFP4 group sizes
#19091
opened Sep 11, 2026 by
moraxu
Collaborator
Loading…
1 task done
[https://nvbugs/6758594][fix] Set
batch_wait_timeout_ms=5000 on the test's LLM — the queue-side knob…
#19090
opened Sep 11, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
[None][test] Supply auxiliary streams in Kimi K3 NVFP4 regression
#19088
opened Sep 11, 2026 by
chang-l
Collaborator
Loading…
[None][fix] drop the placeholder MicroBatchScheduler that shadows the ABC
#19087
opened Sep 11, 2026 by
Anai-Guo
Loading…
3 tasks done
[https://nvbugs/6756848][fix] Exchange the handle as its raw
CUDA_IPC_HANDLE_SIZE (64-byte) struct bytes…
#19086
opened Sep 11, 2026 by
trtllm-agent
Collaborator
Loading…
2 tasks done
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.