1

Is there a way to -in the background- collect cumulative cpu time stats per shared library? Without modifying applications or libraries? Ideally a solution that'll (also) work on arm (Jetson / Pi).

There's /proc/{process_id}/stat to do something at a process level, but is it also possible at object / library file level?

1 Answer 1

0

It's possible but requires support from the kernel and debug symbols for your application and libraries:

$ perf record --all-user --pid PID

# wait at least 1 minute, then press Ctrl-C

$ perf report > perf.log

Without kernel support and debug symbols I'm not sure you can get this information.

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.