I've generally seen time.perf_counter() recommended as the preferred method for measuring time during benchmarking (because time.time() can have pretty bad accuracy on some systems), but this repo uses time.time(). Is there a particular reason for that?
I've generally seen
time.perf_counter()
recommended as the preferred method for measuring time during benchmarking (becausetime.time()
can have pretty bad accuracy on some systems), but this repo usestime.time()
. Is there a particular reason for that?