Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1430662

Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and SPECjvm2008 benchmarks against 4.6 kernel

From Peter Zijlstra <peterz@infradead.org>
Newsgroups linux.kernel
Subject Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and SPECjvm2008 benchmarks against 4.6 kernel
Date 2016-06-24 15:40 +0200
Message-ID <rNzuV-cP-9@gated-at.bofh.it> (permalink)
References (5 earlier) <rNu2d-5cW-7@gated-at.bofh.it> <rNy5P-7Tc-5@gated-at.bofh.it> <rNyIx-86a-13@gated-at.bofh.it> <rNz1T-8uF-21@gated-at.bofh.it> <rNzlf-9y-13@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw

On Fri, Jun 24, 2016 at 03:23:37PM +0200, Vincent Guittot wrote:
> On 24 June 2016 at 15:09, Peter Zijlstra <peterz@infradead.org> wrote:
> > On Fri, Jun 24, 2016 at 02:44:07PM +0200, Vincent Guittot wrote:
> >
> >> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> >> > index 22d64b3f5876..d4f6fb2f3057 100644
> >> > --- a/kernel/sched/fair.c
> >> > +++ b/kernel/sched/fair.c
> >> > @@ -2484,7 +2484,7 @@ static inline long calc_tg_weight(struct task_group *tg, struct cfs_rq *cfs_rq)
> >> >          */
> >> >         tg_weight = atomic_long_read(&tg->load_avg);
> >> >         tg_weight -= cfs_rq->tg_load_avg_contrib;
> >> > -       tg_weight += cfs_rq->load.weight;
> >> > +       tg_weight += cfs_rq->avg.load_avg;
> >>
> >> IIUC, you are reverting
> >> commit  fde7d22e01aa (sched/fair: Fix overly small weight for
> >> interactive group entities)
> >
> > Ah!, I hadn't yet done a git-blame on this. Right you are, we should
> > have put a comment there.
> >
> > So the problem here is that since commit:
> >
> >   2159197d6677 ("sched/core: Enable increased load resolution on 64-bit kernels")
> >
> > load.weight and avg.load_avg are in different metrics. Which completely
> > wrecked things.
> >
> > The obvious alternative is using:
> >
> >         scale_load_down(cfs_rq->load.weight);
> >
> > Let me go run that through the benchmark.
> 
> Yes, looks to be good alternative

Does indeed also work. Let me go write a Changelog and try and magic it
into sched/urgent.

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-23 20:40 +0200
  Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-23 20:50 +0200
    Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Jirka Hladky <jhladky@redhat.com> - 2016-06-24 09:50 +0200
      Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 10:10 +0200
        Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Jirka Hladky <jhladky@redhat.com> - 2016-06-24 10:30 +0200
      Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Jirka Hladky <jhladky@redhat.com> - 2016-06-24 14:10 +0200
        Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Jirka Hladky <jhladky@redhat.com> - 2016-06-24 14:40 +0200
        Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 14:40 +0200
      Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 14:10 +0200
        Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Vincent Guittot <vincent.guittot@linaro.org> - 2016-06-24 14:50 +0200
          Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 15:10 +0200
            Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Vincent Guittot <vincent.guittot@linaro.org> - 2016-06-24 15:30 +0200
              Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 15:40 +0200
              Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 15:50 +0200
          Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Jirka Hladky <jhladky@redhat.com> - 2016-06-24 15:10 +0200
          Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 15:50 +0200
            Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Peter Zijlstra <peterz@infradead.org> - 2016-06-24 18:00 +0200
              Re: Kernel 4.7rc3 - Performance drop 30-40% for SPECjbb2005 and  SPECjvm2008 benchmarks against 4.6 kernel Jirka Hladky <jhladky@redhat.com> - 2016-06-25 00:20 +0200

csiph-web