[pull] main from NVIDIA:main#1474

Merged

pull[bot] merged 21 commits into

LarryXFly:mainfrom

Jul 13, 2026

Commits on Jul 13, 2026

[None][fix] Fix indexer-k-cache transfer for disagg (#16197)
Tabrizian
authored
[None][fix] Correct RocketKV KT cache byte accounting for FP8 KV cache (#16225)
eopXD
authored
[None][perf] serve: opt-in msgspec msgpack transport for disagg orchestrator->worker request body (#16126)
Tabrizian
authored
[None][feat] Add an opt-in raw-weight cache to the HF weight loader (#16054)
sunnyqgg
authored
[https://nvbugs/6430674][fix] Surgical revert of PR #15838's dispatch guard removal: re-add… (#16167)
chenfeiz0326
authored
[None][fix] source DSA metadata from sparse params (#16236)
liji-nv
authored
[None][infra] Waive 7 failed cases for main in post-merge 2835 (#16292)
trtllm-agent
authored
[https://nvbugs/6323074][fix] fix flaky hang issue for disagg gen-onl… (#16172)
bo-nv
authored
[None][test] Waive 9 failed cases for main in QA CI (#16271)
trtllm-agent
authored
[None][infra] Check in most recent lock file from nightly pipeline
tensorrt-cicd
committed
[None][test] Waive 6 failed cases for main in QA CI (#16295)
trtllm-agent
authored
[None][test] Waive 4 failed cases for main in QA CI (#16296)
trtllm-agent
authored
[None][test] Waive 1 failed cases for main in QA CI (#16283)

trtllm-agent
and
jieli-matrix
authored
[https://nvbugs/6438658][fix] Avoid false disagg benchmark KV exhaustion (#16248)
chienchunhung
authored
[https://nvbugs/6330273][fix] Reserve worst-case SWA slots to avoid single-request deadlock (#15588)

lowsfer
and
claude
authored
[None][feat] add per-model KV cache manager v2 auto selection (#15823)
yizhang-nv
authored
[TRTLLM-14138][perf] Make FlashInfer decode plans sync-free with host-built page tables (#16073)
kaiyux
authored
[None][fix] Fix fused mHC output reuse and extend compressor next_n (#16221)
heyuhhh
authored
[https://nvbugs/6428113][fix] Fix TEP token-count handling in MoE Test (#16293)
guqiqi
authored
[None][feat] Support externally provided MPI sessions with explicit ownership (#16053)
sunnyqgg
authored
[None][perf] offload chat template rendering into async (#15284)
yechank-nvidia
authored