[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] xenpm and scheduler



On Tue, Apr 12, 2016 at 11:50:42AM +0000, tutu sky wrote:
> Thanks Dario,
> Yeah, I do like playing with xenpm and try understanding relationship between 
> this and scheduler. but i established a xen environment on vmware, but xenpm 
> does not work correctly and even for 'cpufreq', it is silent at all! so it 
> does not let me try to play with :(. I asked this problem before in the user 
> and devel lists both,  but no body answered me. how can i track the problem, 
> from who (I know that power management is out of your maintenance scope)?? 
> (may xenpm have the same problem on a real platform (again) instead of 
> vmware?)
> thanks a lot.

You will have to.

I don't believe VMWare exposes C and P states to guests so therefore
there is no power freqeuency in play.

> regards.
> ________________________________________
> From: Dario Faggioli <dario.faggioli@xxxxxxxxxx>
> Sent: Tuesday, April 12, 2016 8:10 AM
> To: tutu sky; Meng Xu
> Cc: Xen-devel@xxxxxxxxxxxxx
> Subject: Re: [Xen-devel] xenpm and scheduler
> 
> On Tue, 2016-04-12 at 03:52 +0000, tutu sky wrote:
> > Thanks Xu. I will do as desired about cross messaging.
> >
> > i need it because i exactly want to know that which part of the
> > scheduler's corde (credit), takes effect from this feature. because
> > it is important to me knowing that where would be trade off between
> > idle state and doing load balancing, while cpuidle feature is
> > activated. in other side it's important again for me that what will
> > happen for 'cap; and 'weight' decreasing in a case that one core's
> > frequency is lower than another one in the same socket (actually when
> > cpufreq feature is enable).
> >
> Currently, there is no interaction between the scheduler and the power
> management and frequency scaling layers.
> 
> > Am i clear enough? can you give me an answer or maybe some lines of
> > schedule.c or sched_credit.c's code which i can track them to notice
> > the effect of xenpm on scheduler part of the view?
> >
> If you're saying that, for instance, the CPUs changing frequency can or
> should affect some aspects of the scheduling algorithms (like credits
> burning rate in Credit1 and Credit2, and budget burning rate in RTDS),
> that is an interesting point which may indeed make sense, or at least
> would deserve more investigation.
> 
> But again, right now, there's no line of code to read to understand the
> relationship, as there's no relationship at all.
> 
> If you want to experiment on playing with xenpm, and seeing what effect
> it has on scheduling, that will be very welcome. :-)
> 
> Regards,
> Dario
> --
> <<This happens because I choose it to happen!>> (Raistlin Majere)
> -----------------------------------------------------------------
> Dario Faggioli, Ph.D, http://about.me/dario.faggioli
> Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx
> http://lists.xen.org/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.