merge pr165 commits for Unoptimized library implementation causing CUDA API slow#166

Open

maverick123123 wants to merge 11 commits intoProject-HAMi:mainfrom

maverick123123:fix/rate_limiter

Commits on Mar 30, 2026

Merge pull request #163 from maverick123123/testpr155

hami-robot[bot]
authored and
maverick123123
committed
perf: cache log level to eliminate per-call getenv()/atoi() overhead

nishitnshah
authored and
maverick123123
committed
perf: use cached my_slot pointer in wait_status_self()

nishitnshah
authored and
maverick123123
committed
perf: replace time() syscall with vDSO clock and add double-checked locking in pre_launch_kernel()

nishitnshah
authored and
maverick123123
committed
perf: optimize rate_limiter() by caching limits and removing redundant shared memory ops

nishitnshah
authored and
maverick123123
committed
perf: remove dead cuDeviceGetCount call from oom_check()

nishitnshah
authored and
maverick123123
committed
docs: add performance optimization documentation

nishitnshah
authored and
maverick123123
committed
fix: move log level globals to separate log_utils.c for link visibility

nishitnshah
authored and
maverick123123
committed
fix: restore unconditional set_recent_kernel(2) in rate_limiter()

nishitnshah
authored and
maverick123123
committed
Revert "fix: restore unconditional set_recent_kernel(2) in rate_limiter()"

nishitnshah
authored and
maverick123123
committed
fix: restore set_recent_kernel(2) in rate_limiter() to match original behavior

nishitnshah
authored and
maverick123123
committed