[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <82a781785fd7933d5fa0317de4ba6fe6081c7be0.camel@linux.intel.com>
Date: Mon, 04 Oct 2021 16:06:58 -0700
From: Tim Chen <tim.c.chen@...ux.intel.com>
To: Vincent Guittot <vincent.guittot@...aro.org>, mingo@...hat.com,
peterz@...radead.org, juri.lelli@...hat.com,
dietmar.eggemann@....com, rostedt@...dmis.org, bsegall@...gle.com,
mgorman@...e.de, bristot@...hat.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/2] sched/fair: Improve cost accounting of
newidle_balance
On Mon, 2021-10-04 at 19:14 +0200, Vincent Guittot wrote:
> Improve the precision of the cost of newidle_balance by taking into
> account
> the time spent in updating blocked load and skipping load balance
> loop
> entirely if there is no chance to run at least one.
>
Thanks Vincent.
I'll ask our benchmark team to give it a spin. They've been fairly
busy. I hope to get some bandwidth from them in next couple of
weeks.
Tim
> Vincent Guittot (2):
> sched/fair: account update_blocked_averages in newidle_balance cost
> sched/fair: Skip update_blocked_averages if we are defering load
> balance
>
> kernel/sched/fair.c | 16 +++++++++++-----
> 1 file changed, 11 insertions(+), 5 deletions(-)
>
Powered by blists - more mailing lists