Re: Adding a way to disable the stat cache

Am 03.09.2021 um 18:12 schrieb Kevin Lyda <kevin@lyda.ie <mailto:kevin@lyda.ie>>: > To run this php script: > > <?php > $iterations = 1000000; > function all_the_stats($filename) { > @lstat($filename); > @stat($filename); > } > while ($iterations--) { > all_the_stats(__FILE__); > } > > I see this output: > > Without cache > > real 0m7.326s > user 0m5.877s > sys 0m1.448s > With cache > > real 0m5.010s > user 0m5.009s > sys 0m0.000s So this is almost a 50% performance regression ;-) And the more interesting number for me here is "sys 0m1.448s" vs "sys 0m0.000s". Which means 1.5s out of the additional 2.3s are spent in system calls. Side-note: Why is the user time without cache higher than with cache? That seems counter-intuitive. Maybe the checking of the ini-settings or some libc-code? I'm interested in the load put on a system with a high request count and a typical application. Reducing system calls used to matter there as the kernel does not multi-process the same way user land does. But then again, maybe I'm overly cautious :-) Regards, - Chris