Re: Regression in workingset_refault latency on 5.15

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Mar 2, 2022 at 2:33 PM Ivan Babrou <ivan@xxxxxxxxxxxxxx> wrote:
>
> On Tue, Mar 1, 2022 at 7:40 PM Ivan Babrou <ivan@xxxxxxxxxxxxxx> wrote:
> >
> > On Tue, Mar 1, 2022 at 6:50 PM Shakeel Butt <shakeelb@xxxxxxxxxx> wrote:
> > >
> > > On Tue, Mar 01, 2022 at 04:48:00PM -0800, Ivan Babrou wrote:
> > >
> > > [...]
> > >
> > > > Looks like you were right that targeted flush is not going to be as good.
> > >
> > >
> > > Thanks a lot. Can you please try the following patch (independent of other
> > > patches) as well?
> >
> > Sure thing, I'll come back with the numbers tomorrow.
>
> Measurements from different configurations:
[...]
> avg = 403261 nsecs, total: 30037729036 nsecs, count: 74487
>
> Looks like there are plenty of longer pauses still.

Thanks for running this.

>
> I can also get the numbers from 5.10 for you if you want.

Yes please, that would be good to compare with. Also I am going to
send a new patch in a couple of minutes for you to try and you don't
need to run the "Targeted flush" and the last one which reduces the
memcg_vmstats size.




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux