[None][perf] Cut per-iteration executor bookkeeping in the hang detector and profiler - #17531

[None][perf] Cut per-iteration executor bookkeeping in the hang detector and profiler#17531

pranav-nvidia wants to merge 3 commits into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

pranav-nvidia:perf-executor-iteration-overheadpranav-nvidia/TensorRT-LLM:perf-executor-iteration-overheadCopy head branch name to clipboard