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

Groups > linux.kernel > #1376626

Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup table

Show all headers | View raw

On 12/04/16 03:12, Yuyang Du wrote:
> On Mon, Apr 11, 2016 at 11:41:28AM +0100, Juri Lelli wrote:
> > Hi,
> > 
> > On 11/04/16 06:36, Yuyang Du wrote:
> > > __compute_runnable_contrib() uses a loop to compute sum, whereas a
> > > table loopup can do it faster in a constant time.
> > > 
> > > The following python script can be used to generate the constants:
> > > 
> > > print " #:     yN_inv   yN_sum"
> > > print "-----------------------"
> > > y = (0.5)**(1/32.0)
> > > x = 2**32
> > > xx = 1024
> > > for i in range(0, 32):
> > > 	if i == 0:
> > > 		x = x-1
> > > 		xx = xx*y
> > > 	else:
> > > 		x = x*y
> > > 		xx = int(xx*y + 1024*y)
> > > 	print "%2d: %#x %8d" % (i, int(x), int(xx))
> > > 
> > > print " #:  sum_N32"
> > > print "------------"
> > > xxx = xx
> > > for i in range(0, 11):
> > > 	if i == 0:
> > > 		xxx = xx
> > > 	else:
> > > 		xxx = xxx/2 + xx
> > > 	print "%2d: %8d" % (i, xxx)
> > > 
> > 
> > Thanks for the script, really useful. Do you think there is value in
> > making it general? Like if we want to play with/need changing LOAD_AVG_
> > PERIOD in the future to something different than 32.
> 
> i think a s/32/xx/ should work.
>  
> > Also, does the following assume LOAD_AVG_PERIOD == 32? And if yes, do
> > you think there is any value in removing that assumption?
>  
> Like Peter said, we are heavily dependent on it already.

But I think the current code should still work if we define LOAD_AVG_
PERIOD as, say, 16 and we use Paul's program to recompute the tables.

My point was about trying to keep everything related to LOAD_AVG_PERIOD
and not start assuming it is 32. I'm not saying your changes assume
that, I was asking if they do.

> Whether a half-life
> of 32 periods (or ~32ms) is the best, maybe we can try 16, but definitely not
> 64. Or whether exponential decay is the best to compute the impact of old
> runnable/running times as a pridiction, it is just I can't think of a better
> approach yet, and credits to Paul, Ben, et al.
> 

That is fine, I think. Another thing is crafting the code around a
particular half-life, IMHO.

Best,

- Juri

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

Thread

[PATCH 1/4] sched/fair: Optimize sum computation with a lookup table Yuyang Du <yuyang.du@intel.com> - 2016-04-11 08:20 +0200
  Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup table Vincent Guittot <vincent.guittot@linaro.org> - 2016-04-11 11:10 +0200
  Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Juri Lelli <juri.lelli@arm.com> - 2016-04-11 12:50 +0200
    Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Yuyang Du <yuyang.du@intel.com> - 2016-04-12 05:00 +0200
      Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Juri Lelli <juri.lelli@arm.com> - 2016-04-12 12:20 +0200
        Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Yuyang Du <yuyang.du@intel.com> - 2016-04-13 04:00 +0200
          Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Juri Lelli <juri.lelli@arm.com> - 2016-04-13 11:10 +0200
  Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Dietmar Eggemann <dietmar.eggemann@arm.com> - 2016-04-11 19:00 +0200
    Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Joe Perches <joe@perches.com> - 2016-04-12 01:30 +0200
      Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Juri Lelli <juri.lelli@arm.com> - 2016-04-12 14:10 +0200
    Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Yuyang Du <yuyang.du@intel.com> - 2016-04-12 05:00 +0200
      Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Peter Zijlstra <peterz@infradead.org> - 2016-04-12 16:30 +0200
        Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Yuyang Du <yuyang.du@intel.com> - 2016-04-13 04:00 +0200
  Re: [PATCH 1/4] sched/fair: Optimize sum computation with a lookup  table Joe Perches <joe@perches.com> - 2016-04-12 01:10 +0200

csiph-web