[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Z9i6-_R_LLqt8mq4@slm.duckdns.org>
Date: Mon, 17 Mar 2025 14:14:51 -1000
From: Tejun Heo <tj@...nel.org>
To: Libo Chen <libo.chen@...cle.com>
Cc: Joel Fernandes <joelagnelf@...dia.com>, linux-kernel@...r.kernel.org,
David Vernet <void@...ifault.com>, Andrea Righi <arighi@...dia.com>,
Changwoo Min <changwoo@...lia.com>, Ingo Molnar <mingo@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Juri Lelli <juri.lelli@...hat.com>,
Vincent Guittot <vincent.guittot@...aro.org>,
Dietmar Eggemann <dietmar.eggemann@....com>,
Steven Rostedt <rostedt@...dmis.org>,
Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
Valentin Schneider <vschneid@...hat.com>
Subject: Re: [PATCH RFC] sched_ext: Choose prev_cpu if idle and cache affine
without WF_SYNC
On Mon, Mar 17, 2025 at 05:12:25PM -0700, Libo Chen wrote:
> Just as Peter said, this is whole wake affinity thing is highly workload
> dependent. We have seen cases where even if there are idle cores in the
> prev node, it's still beneficial to cross the NUMA boundary. Will it make
> more sense to have the BPF scheduler implement/alter some of logic here so
> it can fit to different workload?
Oh yeah, multiple SCX schedulers already implement their own select_cpu().
This discussion is just around the default implementation which can be used
if a BPF scheduler doesn't want to implement its own.
Thanks.
--
tejun
Powered by blists - more mailing lists