[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <85115845-93c2-9967-c3b1-749b62b22868@amd.com>
Date: Fri, 15 Sep 2023 08:51:38 +0530
From: K Prateek Nayak <kprateek.nayak@....com>
To: Chen Yu <yu.c.chen@...el.com>
Cc: Tim Chen <tim.c.chen@...el.com>, Aaron Lu <aaron.lu@...el.com>,
Dietmar Eggemann <dietmar.eggemann@....com>,
Steven Rostedt <rostedt@...dmis.org>,
Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
Daniel Bristot de Oliveira <bristot@...hat.com>,
Valentin Schneider <vschneid@...hat.com>,
"Gautham R . Shenoy" <gautham.shenoy@....com>,
linux-kernel@...r.kernel.org,
Peter Zijlstra <peterz@...radead.org>,
Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
Ingo Molnar <mingo@...hat.com>,
Vincent Guittot <vincent.guittot@...aro.org>,
Juri Lelli <juri.lelli@...hat.com>
Subject: Re: [RFC PATCH 2/2] sched/fair: skip the cache hot CPU in
select_idle_cpu()
Hello Chenyu,
Thank you for taking a look at the report.
On 9/14/2023 4:31 PM, Chen Yu wrote:
> Hi Prateek,
>
> thanks for the test,
>
> On 2023-09-14 at 09:43:52 +0530, K Prateek Nayak wrote:
>> [..snip..]
>>
>> Meanwhile, here is result for the current series without any
>> modifications:
>>
>> tl;dr
>>
>> - There seems to be a noticeable increase in hackbench runtime with a
>> single group but big gains beyond that. The regression could possibly
>> be because of added searching but let me do some digging to confirm
>> that.
>
> Ah OK. May I have the command to run 1 group hackbench?
This is actually perf bench sched messaging. The cmdline from the runner
is:
$ perf bench sched messaging -t -p -l 100000 -g <# of groups>
> [..snip..]
--
Thanks and Regards,
Prateek
Powered by blists - more mailing lists