From: Chen Yu <yu.c.chen@intel.com>
To: K Prateek Nayak <kprateek.nayak@amd.com>
Cc: Peter Zijlstra <peterz@infradead.org>, <mingo@redhat.com>,
<juri.lelli@redhat.com>, <vincent.guittot@linaro.org>,
<dietmar.eggemann@arm.com>, <rostedt@goodmis.org>,
<bsegall@google.com>, <mgorman@suse.de>, <bristot@redhat.com>,
<vschneid@redhat.com>, <linux-kernel@vger.kernel.org>,
<wuyun.abel@bytedance.com>, <tglx@linutronix.de>, <efault@gmx.de>,
<tim.c.chen@intel.com>, <yu.c.chen.y@gmail.com>
Subject: Re: [RFC][PATCH 10/10] sched/eevdf: Use sched_attr::sched_runtime to set request/slice suggestion
Date: Tue, 14 May 2024 17:18:55 +0800 [thread overview]
Message-ID: <ZkMsf4Fz7/AFoQfC@chenyu5-mobl2> (raw)
In-Reply-To: <422fc38c-6096-8804-17ce-1420661743e8@amd.com>
Hi Prateek,
On 2024-05-13 at 09:37:07 +0530, K Prateek Nayak wrote:
> Hello Chenyu,
>
> On 5/7/2024 8:45 PM, Chen Yu wrote:
> > On 2024-04-05 at 12:28:04 +0200, Peter Zijlstra wrote:
> >> Allow applications to directly set a suggested request/slice length using
> >> sched_attr::sched_runtime.
> >>
> >> The implementation clamps the value to: 0.1[ms] <= slice <= 100[ms]
> >> which is 1/10 the size of HZ=1000 and 10 times the size of HZ=100.
> >>
> >> Applications should strive to use their periodic runtime at a high
> >> confidence interval (95%+) as the target slice. Using a smaller slice
> >> will introduce undue preemptions, while using a larger value will
> >> increase latency.
> >>
> >> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> >>
> >
> > Is it possible to leverage this task slice to do better task wakeup placement?
> > The idea is that, the smaller the slice the wakee has, the less idle CPU it
> > should scan. This can reduce wake latency and inhibit costly task migration,
> > especially on large systems.
> >
> > We did some experiments and got some performance improvements:
> >
> >
> > From 9cb806476586d7048fcbd0f66d0101f0dbb8fd2b Mon Sep 17 00:00:00 2001
> > From: Chen Yu <yu.c.chen@intel.com>
> > Date: Tue, 7 May 2024 22:36:29 +0800
> > Subject: [RFC PATCH] sched/eevdf: Use customized slice to reduce wakeup latency
> > and inhibit task migration
> >
> > Problem 1:
> > The overhead of task migration is high on many-core system. The overhead
> > brings performance penalty due to broken cache locality/higher cache-to-cache
> > latency.
> >
> > Problem 2:
> > During wakeup, the time spent on searching for an idle CPU is costly on
> > many-core system. Besides, access to other CPU's rq statistics brings
> > cace contention:
> >
> > available_idle_cpu(cpu) -> idle_cpu(cpu) -> {rq->curr, rq->nr_running}
> >
> > Although SIS_UTIL throttles the scan depth based on system utilization,
> > there is requirement to further limit the scan depth for specific workload,
> > especially for short duration wakee.
> >
> > Now we have the interface to customize the request/slice. The smaller the
> > slice is, the earlier the task can be picked up, and the lower wakeup latency
> > the task expects. Leverage the wakee's slice to further throttle the
> > idle CPU scan depth - the shorter slice, the less CPUs to scan.
> >
> > Test on 240 CPUs, 2 sockets system. With SNC(sub-numa-cluster) enabled,
> > each LLC domain has 60 CPUs. There is noticeable improvement of netperf.
> > (With SNC disabled, more improvements should be seen because C2C is higher)
> >
> > The global slice is 3 msec(sysctl_sched_base_slice) by default on my ubuntu
> > 22.04, and the customized slice is set to 0.1 msec for both netperf and netserver:
> >
> > for i in $(seq 1 $job); do
> > netperf_slice -e 100000 -4 -H 127.0.01 -t TCP_RR -c -C -l 100 &
> > done
> >
> > case load baseline(std%) compare%( std%)
> > TCP_RR 60-threads 1.00 ( 1.60) +0.35 ( 1.73)
> > TCP_RR 120-threads 1.00 ( 1.34) -0.96 ( 1.24)
> > TCP_RR 180-threads 1.00 ( 1.59) +92.20 ( 4.24)
> > TCP_RR 240-threads 1.00 ( 9.71) +43.11 ( 2.97)
> >
> > Suggested-by: Tim Chen <tim.c.chen@intel.com>
> > Signed-off-by: Chen Yu <yu.c.chen@intel.com>
> > ---
> > kernel/sched/fair.c | 23 ++++++++++++++++++++---
> > kernel/sched/features.h | 1 +
> > 2 files changed, 21 insertions(+), 3 deletions(-)
> >
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index edc23f6588a3..f269ae7d6e24 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -7368,6 +7368,24 @@ static inline int select_idle_smt(struct task_struct *p, struct sched_domain *sd
> >
> > #endif /* CONFIG_SCHED_SMT */
> >
> > +/*
> > + * Scale the scan number of idle CPUs according to customized
> > + * wakee's slice. The smaller the slice is, the earlier the task
> > + * wants be picked up, thus the lower wakeup latency the task expects.
> > + * The baseline is the global sysctl_sched_base_slice. Task slice
> > + * smaller than the global one would shrink the scan number.
> > + */
> > +static int adjust_idle_scan(struct task_struct *p, int nr)
> > +{
> > + if (!sched_feat(SIS_FAST))
> > + return nr;
> > +
> > + if (!p->se.custom_slice || p->se.slice >= sysctl_sched_base_slice)
> > + return nr;
> > +
> > + return div_u64(nr * p->se.slice, sysctl_sched_base_slice);
> > +}
> > +
> > /*
> > * Scan the LLC domain for idle CPUs; this is dynamically regulated by
> > * comparing the average scan cost (tracked in sd->avg_scan_cost) against the
> > @@ -7384,10 +7402,9 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, bool
> > if (sched_feat(SIS_UTIL)) {
> > sd_share = rcu_dereference(per_cpu(sd_llc_shared, target));
> > if (sd_share) {
> > - /* because !--nr is the condition to stop scan */> - nr = READ_ONCE(sd_share->nr_idle_scan) + 1;
> > + nr = adjust_idle_scan(p, READ_ONCE(sd_share->nr_idle_scan));
> > /* overloaded LLC is unlikely to have idle cpu/core */
> > - if (nr == 1)
> > + if (nr <= 0)
>
> I was wondering if this would preserve the current behavior with
> SIS_FAST toggled off? Since the implementation below still does a
> "--nr <= 0" , wouldn't it effectively visit one CPU less overall now?
>
> Have you tried something similar to the below hunk?
>
> /* because !--nr is the condition to stop scan */
> nr = adjust_idle_scan(p, READ_ONCE(sd_share->nr_idle_scan)) + 1;
> if (nr == 1)
> return -1;
>
Yeah, right, to keep the scan depth consistent, the "+1" should be kept.
> I agree with Mike that looking at slice to limit scan-depth seems odd.
> My experience with netperf is that the workload cares more about the
> server-client being co-located on the closest cache domain and by
> limiting scan-depth using slice, this is indirectly achieved since all
> the wakeups carry the WF_SYNc flag.
>
Exactly. This is the original motivation.
> P.S. have you tried using the slice in __select_idle_cpu()? Similar to
> sched_idle_cpu() check, perhaps an additional sched_preempt_short_cpu()
> which compares rq->curr->se.slice with the waking task's slice and
> returs that cpu if SIS_SHORT can help run the workload quicker?
This is a good idea, it seems to be benefit PREEMPT_SHORT. If the customized
task slice is introduced, we can leverage this hint for latency related
optimization. Task wakeup is one thing, I can also think of other aspects,
like idle load balance, etc. I'm not sure what is the proper usage of the
task slice though, this is why I sent this RFC.
> Note:
> This will not work if the SIS scan itself is the largest overhead in the
> wakeup cycle and not the task placement itself. Previously during
> SIS_UTIL testing, to measure the overheads of scan vs placement, we
> would do a full scan but return the result that SIS_UTIL would have
> returned to determine the overhead of the search itself.
>
Regarding the task placement, do you mean the time between a task is enqueued
and picked up? Do you have any recommendation which workload can expose the
scan overhead most?
thanks,
Chenyu
> > return -1;
> > }
> > }
> > diff --git a/kernel/sched/features.h b/kernel/sched/features.h
> > index 143f55df890b..176324236018 100644
> > --- a/kernel/sched/features.h
> > +++ b/kernel/sched/features.h
> > @@ -50,6 +50,7 @@ SCHED_FEAT(TTWU_QUEUE, true)
> > * When doing wakeups, attempt to limit superfluous scans of the LLC domain.
> > */
> > SCHED_FEAT(SIS_UTIL, true)
> > +SCHED_FEAT(SIS_FAST, true)
> >
> > /*
> > * Issue a WARN when we do multiple update_rq_clock() calls
>
> --
> Thanks and Regards,
> Prateek
next prev parent reply other threads:[~2024-05-14 9:19 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-05 10:27 [RFC][PATCH 00/10] sched/fair: Complete EEVDF Peter Zijlstra
2024-04-05 10:27 ` [RFC][PATCH 01/10] sched/eevdf: Add feature comments Peter Zijlstra
2024-04-05 10:27 ` [RFC][PATCH 02/10] sched/eevdf: Remove min_vruntime_copy Peter Zijlstra
2024-04-05 10:27 ` [RFC][PATCH 03/10] sched/fair: Cleanup pick_task_fair() vs throttle Peter Zijlstra
2024-04-05 21:11 ` Benjamin Segall
2024-04-05 10:27 ` [RFC][PATCH 04/10] sched/fair: Cleanup pick_task_fair()s curr Peter Zijlstra
2024-04-05 10:27 ` [RFC][PATCH 05/10] sched/fair: Unify pick_{,next_}_task_fair() Peter Zijlstra
2024-04-06 2:20 ` Mike Galbraith
2024-04-05 10:28 ` [RFC][PATCH 06/10] sched: Allow sched_class::dequeue_task() to fail Peter Zijlstra
2024-04-05 10:28 ` [RFC][PATCH 07/10] sched/fair: Re-organize dequeue_task_fair() Peter Zijlstra
2024-04-05 10:28 ` [RFC][PATCH 08/10] sched/fair: Implement delayed dequeue Peter Zijlstra
2024-04-06 9:23 ` Chen Yu
2024-04-08 9:06 ` Peter Zijlstra
2024-04-11 1:32 ` Yan-Jie Wang
2024-04-25 10:25 ` Peter Zijlstra
2024-04-12 10:42 ` K Prateek Nayak
2024-04-15 10:56 ` Mike Galbraith
2024-04-16 3:18 ` K Prateek Nayak
2024-04-16 5:36 ` Mike Galbraith
2024-04-18 16:24 ` Mike Galbraith
2024-04-18 17:08 ` K Prateek Nayak
2024-04-24 15:20 ` Peter Zijlstra
2024-04-25 11:28 ` Peter Zijlstra
2024-04-26 10:56 ` Peter Zijlstra
2024-04-26 11:16 ` Peter Zijlstra
2024-04-26 16:03 ` Mike Galbraith
2024-04-27 6:42 ` Mike Galbraith
2024-04-28 16:32 ` Mike Galbraith
2024-04-29 12:14 ` Peter Zijlstra
2024-04-15 17:07 ` Luis Machado
2024-04-24 15:15 ` Luis Machado
2024-04-25 10:42 ` Peter Zijlstra
2024-04-25 11:49 ` Peter Zijlstra
2024-04-26 9:32 ` Peter Zijlstra
2024-04-26 9:36 ` Peter Zijlstra
2024-04-26 10:16 ` Luis Machado
2024-04-29 14:33 ` Luis Machado
2024-05-02 10:26 ` Luis Machado
2024-05-10 14:49 ` Luis Machado
2024-05-15 9:36 ` Peter Zijlstra
2024-05-15 11:48 ` Peter Zijlstra
2024-05-15 18:03 ` Mike Galbraith
2024-05-20 15:20 ` Luis Machado
2024-05-29 22:50 ` Peter Zijlstra
2024-06-03 19:30 ` Luis Machado
2024-06-04 10:11 ` Peter Zijlstra
2024-06-04 13:59 ` Hongyan Xia
2024-06-04 14:23 ` Luis Machado
2024-06-04 14:49 ` Hongyan Xia
2024-06-04 19:12 ` Peter Zijlstra
2024-06-05 7:22 ` Peter Zijlstra
2024-06-05 9:14 ` Luis Machado
2024-06-05 9:42 ` Peter Zijlstra
2024-06-12 15:08 ` Luis Machado
2024-05-23 8:45 ` Peter Zijlstra
2024-05-23 9:06 ` Luis Machado
2024-05-23 9:33 ` Peter Zijlstra
2024-06-03 15:57 ` Hongyan Xia
2024-04-26 10:15 ` Luis Machado
2024-04-20 5:57 ` Mike Galbraith
2024-04-22 13:13 ` Tobias Huschle
2024-04-05 10:28 ` [RFC][PATCH 09/10] sched/eevdf: Allow shorter slices to wakeup-preempt Peter Zijlstra
2024-04-05 10:28 ` [RFC][PATCH 10/10] sched/eevdf: Use sched_attr::sched_runtime to set request/slice suggestion Peter Zijlstra
2024-04-06 8:16 ` Hillf Danton
2024-05-07 5:34 ` Mike Galbraith
2024-05-15 10:13 ` Peter Zijlstra
2024-05-07 15:15 ` Chen Yu
2024-05-08 13:52 ` Mike Galbraith
2024-05-09 3:48 ` Chen Yu
2024-05-09 5:00 ` Mike Galbraith
2024-05-13 4:07 ` K Prateek Nayak
2024-05-14 9:18 ` Chen Yu [this message]
2024-05-14 15:23 ` K Prateek Nayak
2024-05-14 16:15 ` Chen Yu
2024-05-22 14:48 ` Chen Yu
2024-05-27 10:11 ` [RFC][PATCH 00/10] sched/fair: Complete EEVDF K Prateek Nayak
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZkMsf4Fz7/AFoQfC@chenyu5-mobl2 \
--to=yu.c.chen@intel.com \
--cc=bristot@redhat.com \
--cc=bsegall@google.com \
--cc=dietmar.eggemann@arm.com \
--cc=efault@gmx.de \
--cc=juri.lelli@redhat.com \
--cc=kprateek.nayak@amd.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mgorman@suse.de \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=tim.c.chen@intel.com \
--cc=vincent.guittot@linaro.org \
--cc=vschneid@redhat.com \
--cc=wuyun.abel@bytedance.com \
--cc=yu.c.chen.y@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).