Re: [PATCH v2 10/10] cpufreq: schedutil: New governor based on scheduler utilization data

[Srinivas Pandruvada]

On Fri, 2016-03-04 at 11:26 +0000, Juri Lelli wrote:
> Hi Rafael,
> 
> On 04/03/16 04:35, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
> > 
> > Add a new cpufreq scaling governor, called "schedutil", that uses
> > scheduler-provided CPU utilization information as input for making
> > its decisions.
> > 
> > Doing that is possible after commit fe7034338ba0 (cpufreq: Add
> > mechanism for registering utilization update callbacks) that
> > introduced cpufreq_update_util() called by the scheduler on
> > utilization changes (from CFS) and RT/DL task status updates.
> > In particular, CPU frequency scaling decisions may be based on
> > the the utilization data passed to cpufreq_update_util() by CFS.
> > 
> > The new governor is relatively simple.
> > 
> > The frequency selection formula used by it is
> > 
> > 	next_freq = util * max_freq / max
> > 
> > where util and max are the utilization and CPU capacity coming from
> > CFS.
> > 
> 
> The formula looks better to me now. However, problem is that, if you
> have freq. invariance, util will slowly saturate to the current
> capacity. So, we won't trigger OPP changes for a task that for
> example
> starts light and then becomes big.
> 
> This is the same problem we faced with schedfreq. The current
> solution
> there is to use a margin for calculating a threshold (80% of current
> capacity ATM). Once util goes above that threshold we trigger an OPP
> change.ÂÂCurrent policy is pretty aggressive, we go to max_f and then
> adapt to the "real" util during successive enqueues. This was also
> tought to cope with the fact that PELT seems slow to react to abrupt
> changes in tasks behaviour.
> 
I also tried something like this in intel_pstate with scheduler util,
where you ramp up to turbo when a threshold percent exceeded then ramp
down slowly in steps. This helped some workloads like tbench to perform
better, but it resulted in lower performance/watt on specpower server
workload. The problem is finding what is the right threshold value.

Thanks,
Srinivas

> I'm not saying this is the definitive solution, but I fear something
> along this line is needed when you add freq invariance in the mix.
> 
> Best,
> 
> - Juri
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pm"
> in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info atÂÂhttp://vger.kernel.org/majordomo-info.html