Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 1 Mar 2021 19:07:20 +0530
From:   Srikar Dronamraju <srikar@linux.vnet.ibm.com>
To:     Rik van Riel <riel@surriel.com>
Cc:     Ingo Molnar <mingo@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Mel Gorman <mgorman@techsingularity.net>,
        Thomas Gleixner <tglx@linutronix.de>,
        Valentin Schneider <valentin.schneider@arm.com>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Michael Neuling <mikey@neuling.org>,
        Gautham R Shenoy <ego@linux.vnet.ibm.com>,
        Parth Shah <parth@linux.ibm.com>
Subject: Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity
Message-ID: <20210301133720.GI2028034@linux.vnet.ibm.com>
Reply-To: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
References: <20210226164029.122432-1-srikar@linux.vnet.ibm.com>
 <ab046b9d5bcd29b2eb759cd999e2f578a683c673.camel@surriel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <ab046b9d5bcd29b2eb759cd999e2f578a683c673.camel@surriel.com>
Precedence: bulk

* Rik van Riel <riel@surriel.com> [2021-02-27 14:56:07]:

> > In the current situation where waker and previous CPUs are busy, but
> > only one of its LLC has an idle CPU, Scheduler may end up picking a
> > LLC
> > with no idle CPUs. To mitigate this, add a new step between 1 and 2
> > where Scheduler compares idle CPUs in waker and previous LLCs and
> > picks
> > the appropriate one.
> 
> I like that idea a lot. That could also solve some of the

Thanks.

> issues sometimes observed on multi-node x86 systems, and
> probably on the newer AMD chips with several LLCs on chip.
> 

Okay.

> > +	if (sched_feat(WA_WAKER) && tnr_busy < tllc_size)
> > +		return this_cpu;
> 
> I wonder if we need to use a slightly lower threshold on
> very large LLCs, both to account for the fact that the
> select_idle_cpu code may not find the single idle CPU
> among a dozen busy ones, or because on a system with
> hyperthreading we may often be better off picking another
> LLC for HT contention issues?

> 
> Maybe we could use "tnr_busy * 4 <
> tllc_size * 3" or
> something like that?
> 
> That way we will only try to find the last 5 idle
> CPUs
> in a 22 CPU LLC if the other LLC also has fewer than 6
> idle cores.
> 
> That might increase our chances of finding an idle CPU
> with SIS_PROP enabled, and might allow WA_WAKER to be
> true by default.

Agree we need to be conservative esp if we want to make WA_WAKER on by
default. I would still like to hear from other people if they think its ok
to enable it by default. I wonder if enabling it by default can cause some
load imbalances leading to more active load balance down the line.  I
haven't benchmarked with WA_WAKER enabled.

Thanks Rik for your inputs.

-- 
Thanks and Regards
Srikar Dronamraju