Re: [PATCH 2/2] sched,time: call __acct_update_integrals once a jiffy

[Mike Galbraith]

On Sun, 2016-01-31 at 03:52 +0100, Mike Galbraith wrote:
> On Sat, 2016-01-30 at 21:36 +0100, Frederic Weisbecker wrote:
> > On Sat, Jan 30, 2016 at 06:53:05PM +0100, Mike Galbraith wrote:
> > > On Sat, 2016-01-30 at 15:20 +0100, Frederic Weisbecker wrote:
> > > > On Fri, Jan 29, 2016 at 05:43:28PM -0500, Rik van Riel wrote:
> > > 
> > > > > Run times for the microbenchmark:
> > > > > 
> > > > > 4.4                             3.8 seconds
> > > > > 4.5-rc1                         3.7 seconds
> > > > > 4.5-rc1 + first patch           3.3 seconds
> > > > > 4.5-rc1 + both patches          2.3 seconds
> > > > 
> > > > Very nice improvement!
> > > 
> > > Tasty indeed.
> > > 
> > > When nohz_full CPUs are not isolated, ie are being used as
> > > generic
> > > CPUs, get_nohz_timer_target() is a problem with things like
> > > tbench.
> > 
> > So by isolated CPU you mean those part of isolcpus= boot option,
> > right?
> 
> Yes, isolated in the scheduler sense, either via isolcpus= or
> cpusets. 
>  If the CPU is part of a scheduler domain, it is by definition part
> of
> the generic work crew.
>  
> > > tbench 8 with Rik's patches applied:
> > > nohz_full=empty
> > > Throughput 3204.69 MB/sec  1.000
> > > nohz_full=1-3,5-7 
> > > Throughput 1354.99 MB/sec   .422  1.000
> > > nohz_full=1-3,5-7 + club below 
> > > Throughput 2762.22 MB/sec   .861  2.038
> > > 
> > > With Rik's patches and a club, tbench becomes nearly acceptable.
> > > ---
> > >  include/linux/tick.h |    2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > --- a/include/linux/tick.h
> > > +++ b/include/linux/tick.h
> > > @@ -184,7 +184,7 @@ static inline const struct cpumask *hous
> > >  static inline bool is_housekeeping_cpu(int cpu)
> > >  {
> > >  #ifdef CONFIG_NO_HZ_FULL
> > > -	if (tick_nohz_full_enabled())
> > > +	if (tick_nohz_full_enabled() &&
> > > runqueue_is_isolated(cpu))
> > >  		return cpumask_test_cpu(cpu, housekeeping_mask);
> > 
> > This makes me confused. How forcing timers to CPUs in isolcpus is
> > making
> > better results?
> 
> It doesn't, it's shutting get_nohz_timer_target() down for those
> nohz_full CPUs that are NOT currently isolated, are thus generic CPUs
> with the capability to _become_ elite solo artists on demand.

Damn, I'm gonna have to ask...

If an isolated elite task sets the alarm, why would it want it to go
off somewhere else, maybe in the middle of a death metal concert? 
 Seems it must induce jitter.  Why is this a good thing?

	-Mike