Re: [RFCv3 PATCH 18/48] sched: Track blocked utilization contributions

[Morten Rasmussen]

On Tue, Mar 24, 2015 at 04:07:29PM +0000, Peter Zijlstra wrote:
> On Tue, Mar 24, 2015 at 09:43:47AM +0000, Morten Rasmussen wrote:
> > On Mon, Mar 23, 2015 at 02:08:01PM +0000, Peter Zijlstra wrote:
> > > On Wed, Feb 04, 2015 at 06:30:55PM +0000, Morten Rasmussen wrote:
> > > > Introduces the blocked utilization, the utilization counter-part to
> > > > cfs_rq->utilization_load_avg. It is the sum of sched_entity utilization
> > > > contributions of entities that were recently on the cfs_rq that are
> > > > currently blocked. Combined with sum of contributions of entities
> > > > currently on the cfs_rq or currently running
> > > > (cfs_rq->utilization_load_avg) this can provide a more stable average
> > > > view of the cpu usage.
> > > 
> > > So it would be nice if you add performance numbers for all these patches
> > > that add accounting muck.. 
> > 
> > Total scheduler latency (as in hackbench?), individual function
> > latencies, or something else?
> 
> Yeah, good question that. Something that is good at running this code a
> lot. So dequeue_entity() -> dequeue_entity_load_avg() ->
> update_entity_load_avg() -> __update_entity_runnable_avg() seems a
> reliable way into here, and IIRC hackbench does a lot of that, so yes,
> that might just work.

Hackbench does a lot of that. I used it recently to measure the impact
of the weak arch_scale_*() functions. I will dig out some numbers.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/