[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230815235201.GH971582@hirez.programming.kicks-ass.net>
Date: Wed, 16 Aug 2023 01:52:01 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Mike Galbraith <umgwanakikbuti@...il.com>
Cc: Chen Yu <yu.c.chen@...el.com>,
kernel test robot <oliver.sang@...el.com>,
oe-lkp@...ts.linux.dev, lkp@...el.com,
linux-kernel@...r.kernel.org, x86@...nel.org,
Ingo Molnar <mingo@...nel.org>
Subject: Re: [tip:sched/eevdf] [sched/fair] e0c2ff903c:
phoronix-test-suite.blogbench.Write.final_score -34.8% regression
On Mon, Aug 14, 2023 at 08:32:55PM +0200, Mike Galbraith wrote:
> I don't have that phoronix thingy, but the above didn't seem to do
> anything for hackbench here.
Yeah, looking at it this is because the lag and the request/slice are on
the same order of things.
More thinking required, but perhaps not at 2am.
> I whack eevdf with the stick below to dampen its scheduling enthusiasm
> a little, and it did help the hackbench deficit some.
I'm thinking there's an argument to be had the current WAKEUP_PREEMPTION
knob is at the wrong place and should be more or less where you put the
triddle at.
Powered by blists - more mailing lists