Re: [sched/fair] 38ac256d1c: stress-ng.vm-segv.ops_per_sec -13.8% regression
From: Valentin Schneider
Date: Thu Apr 22 2021 - 16:42:38 EST
On 22/04/21 10:55, Valentin Schneider wrote:
> I'll go find myself some other x86 box and dig into it;
> I'd rather not leave this hanging for too long.
So I found myself a dual-socket Xeon Gold 5120 @ 2.20GHz (64 CPUs) and
*there* I get a somewhat consistent ~-6% regression. As I'm suspecting
cacheline shenanigans, I also ran that with Peter's recent
kthread_is_per_cpu() change, and that brings it down to ~-3%
I'll leave it at here for today, but at least that's something I can work
with.