whisper.cpp
whisper.cpp copied to clipboard
Should we use process time for benchmarks/timings?
Currently, the bench
tool simply spits out the timings from whisper_print_timings
. These timings are not process specific and as such are influenced by other processes on the system.
Should we use something like CLOCK_PROCESS_CPUTIME_ID
instead of CLOCK_MONOTONIC
in ggml_time_*
? This would make the benchmarks in #89 outdated, but would make future benchmarks more realistic.
I think we can measure and report both.
On non-Windows system we can use the existing ggml_cycles()
:
https://github.com/ggerganov/whisper.cpp/blob/d1ea1220ffeff846efd0578e8df3b1d1b802bb5d/ggml.c#L315-L321
Yeah, I did see that and also that they aren't used anywhere other than commented out debugging statements - I'll add a field that includes them in whisper_print_timings
tomorrow.