ggml-org/whisper.cpp
Should we use process time for benchmarks/timings?
Ouverte
#382 ouverte le 7 janv. 2023
enhancementgood first issue
Métriques du dépôt
- Stars
- (52 942 étoiles)
- Métriques de merge PR
- (Merge moyen 5j 20h) (36 PRs mergées en 30 j)
Description
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 This would make the benchmarks in #89 outdated, but would make future benchmarks more realistic.