Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1613929
| From | Peter Zijlstra <peterz@infradead.org> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() |
| Date | 2017-03-31 13:50 +0200 |
| Message-ID | <tr3e2-45U-17@gated-at.bofh.it> (permalink) |
| References | (4 earlier) <tqjkR-5EX-11@gated-at.bofh.it> <tqGr8-5AE-3@gated-at.bofh.it> <tqHdw-69w-15@gated-at.bofh.it> <tqJ5D-7vN-11@gated-at.bofh.it> <tr2rE-3x5-21@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
On Fri, Mar 31, 2017 at 03:55:40AM -0700, Paul Turner wrote:
> > I'm thinking that we can compute the middle segment, by taking the max
> > value and chopping off the ends, like:
> >
>
> >
> > p
> > c2 = 1024 \Sum y^n
> > n=1
> >
> > inf inf
> > = 1024 ( \Sum y^n - \Sum y^n - y^0 )
> > n=0 n=p
> >
>
> Very nice!
> Minor nit: Second sum needs to be from n=p+1
Correct.
> > +static u32 __accumulate_pelt_segments(u64 periods, u32 d1, u32 d3)
> > {
> > + u32 c1, c2, c3 = d3; /* y^0 == 1 */
> >
> > /*
> > * c1 = d1 y^(p+1)
> > */
> > + c1 = decay_load((u64)d1, periods);
> >
> > /*
> > + * p
> > + * c2 = 1024 \Sum y^n
> > + * n=1
> > *
> > + * inf inf
> > + * = 1024 ( \Sum y^n - \Sum y^n - y^0 )
> > + * n=0 n=p+1
> > */
> > + c2 = LOAD_AVG_MAX - decay_load(LOAD_AVG_MAX, periods) - 1024;
>
> decay_load(LOAD_AVG_MAX, periods + 1)
So here, @periods == p+1, see also c1. Yes, this is confusing [*].
In particular, I think the decay terms for c1 and this should be the
same. We cut off this tail end of the series to replace it with c1 after
all.
[*] hysterically p used to be off by 1, which is where the p+1 came
from, but now periods includes it. I was thinking of doing a patch
correcting all the comments to fully eradicate the whole +1 business.
> I computed all the values vs true value that the old/new computations
> result in, and it's very close. Absolutely it's approximately 2x off
> the previous computation, e.g. if the old value was -15 (relative to
> true value) than the new computation is -30.
>
> This is definitely more than good enough. If we want more precision,
> then the correction factor of:
> +clamp(periods, 0, 45)
Can you do a patch with coherent comment explaining where that
correction term comes from?
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-28 16:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Yuyang Du <yuyang.du@intel.com> - 2017-03-29 10:20 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-29 12:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Yuyang Du <yuyang.du@intel.com> - 2017-03-30 04:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Paul Turner <pjt@google.com> - 2017-03-30 13:30 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-30 14:20 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-30 15:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Yuyang Du <yuyang.du@intel.com> - 2017-03-31 05:00 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-30 16:20 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Paul Turner <pjt@google.com> - 2017-03-31 00:10 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-31 09:10 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Paul Turner <pjt@google.com> - 2017-03-31 12:00 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-31 13:30 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Dietmar Eggemann <dietmar.eggemann@arm.com> - 2017-04-10 09:40 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-04-10 10:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-31 13:40 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Yuyang Du <yuyang.du@intel.com> - 2017-03-31 05:20 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Yuyang Du <yuyang.du@intel.com> - 2017-03-31 05:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-31 09:20 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Paul Turner <pjt@google.com> - 2017-03-31 13:00 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-03-31 13:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Peter Zijlstra <peterz@infradead.org> - 2017-04-10 12:50 +0200
Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg() Yuyang Du <yuyang.du@intel.com> - 2017-03-31 04:50 +0200
csiph-web