linux-kernel - Re: [PATCH 1/2] sched/fair: Couple wakee flips with heavy wakers

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <91cb83b8e05076bd6650e3bfcda00feb890b911a.camel@gmx.de>
Date:   Sat, 30 Oct 2021 06:12:17 +0200
From:   Mike Galbraith <efault@....de>
To:     Vincent Guittot <vincent.guittot@...aro.org>,
        Mel Gorman <mgorman@...hsingularity.net>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...nel.org>,
        Valentin Schneider <Valentin.Schneider@....com>,
        Aubrey Li <aubrey.li@...ux.intel.com>,
        Barry Song <song.bao.hua@...ilicon.com>,
        Srikar Dronamraju <srikar@...ux.vnet.ibm.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] sched/fair: Couple wakee flips with heavy wakers

On Sat, 2021-10-30 at 05:11 +0200, Mike Galbraith wrote:
>
> I profiled it with a perf that sums delay (local mod I find useful)...

Here are those numbers for completeness.  As you can see, patchlet was
fairly meaningless to the desktop load.  In fact, even though the wake
affine logic essentially favors the compute load over the heavily
threaded but light desktop load, I noticed zero difference with or
without that being allowed, nor did perf record any noteworthy latency
events, just more wait as load was kinda sorta de-treaded by being
stacked up.

box = i7-4790 quad+smt
desktop vs massive_intr 8 9999 (8 x 8ms run/1ms sleep, for 9999 secs.. effectively forever)
perf sched record -a -- su mikeg -c 'firefox https://www.youtube.com/watch?v=aqz-KE-bpKQ'& sleep 300 && killall perf firefox
                     runtime              runtime               sum delay      sum delay        sum delay       switches      desktop
patch/features       total          util  massive_intr   util   total          massive_intr     desktop        total/massive  util
virgin/stock         2267347.921 ms 94.4% 1932675.152 ms 80.5%  158611.016 ms  133309.938 ms    25301.078 ms  594780/441157   13.9%
virgin/-wa_weight    2236871.408 ms 93.2% 1881464.401 ms 78.3%  255785.391 ms  243958.616 ms    11826.775 ms 1525470/1424083  14.8%
                          -1.34%    -1.2%                -2.2%                                    -13.474 s                   +0.9%
wake_wide/stock      2254335.961 ms 93.9% 1917834.157 ms 79.9%  164766.194 ms  141974.540 ms    22791.654 ms  720711/599064   14.0%