* [Qemu-devel] question about VM kernel parameter idle=<poll/mwait/halt/nomwait>
@ 2013-11-21 3:45 Zhanghaoyu (A)
2013-11-21 7:01 ` Michael S. Tsirkin
0 siblings, 1 reply; 3+ messages in thread
From: Zhanghaoyu (A) @ 2013-11-21 3:45 UTC (permalink / raw)
To: qemu-devel@nongnu.org, KVM, Michael S. Tsirkin, Gleb Natapov,
Paolo Bonzini
Cc: Huangweidong (C), Luonengjun
Hi, all
What's the difference of the linux guest kernel parameter idle=<poll/mwait/halt/nomwait>, especially in performance?
Taking the performance into account, which one is best?
In my opinion, if the number of all VMs' vcpus is far more than that of pcpus, e.g. SPECVirt test, idle=halt is better for server's total throughput,
otherwise, e.g. in some CT scenario, the number of total vcpus is not greater than that of pcpus, idle=poll is better for server's total throughput,
because of less latency and VMEXIT.
linux-3.9 and above, idle=mwait is not recommended.
Thanks,
Zhang Haoyu
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [Qemu-devel] question about VM kernel parameter idle=<poll/mwait/halt/nomwait>
2013-11-21 3:45 [Qemu-devel] question about VM kernel parameter idle=<poll/mwait/halt/nomwait> Zhanghaoyu (A)
@ 2013-11-21 7:01 ` Michael S. Tsirkin
2013-11-21 7:22 ` Gleb Natapov
0 siblings, 1 reply; 3+ messages in thread
From: Michael S. Tsirkin @ 2013-11-21 7:01 UTC (permalink / raw)
To: Zhanghaoyu (A)
Cc: Huangweidong (C), KVM, Gleb Natapov, Luonengjun,
qemu-devel@nongnu.org, Paolo Bonzini
On Thu, Nov 21, 2013 at 03:45:28AM +0000, Zhanghaoyu (A) wrote:
> Hi, all
>
> What's the difference of the linux guest kernel parameter idle=<poll/mwait/halt/nomwait>, especially in performance?
>
> Taking the performance into account, which one is best?
>
> In my opinion, if the number of all VMs' vcpus is far more than that of pcpus, e.g. SPECVirt test, idle=halt is better for server's total throughput,
> otherwise, e.g. in some CT scenario, the number of total vcpus is not greater than that of pcpus, idle=poll is better for server's total throughput,
> because of less latency and VMEXIT.
Makes sense overall.
> linux-3.9 and above, idle=mwait is not recommended.
>
> Thanks,
> Zhang Haoyu
Does it actually have effect? I didn't think it would.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [Qemu-devel] question about VM kernel parameter idle=<poll/mwait/halt/nomwait>
2013-11-21 7:01 ` Michael S. Tsirkin
@ 2013-11-21 7:22 ` Gleb Natapov
0 siblings, 0 replies; 3+ messages in thread
From: Gleb Natapov @ 2013-11-21 7:22 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Huangweidong (C), KVM, Zhanghaoyu (A), Luonengjun,
qemu-devel@nongnu.org, Paolo Bonzini
On Thu, Nov 21, 2013 at 09:01:39AM +0200, Michael S. Tsirkin wrote:
> On Thu, Nov 21, 2013 at 03:45:28AM +0000, Zhanghaoyu (A) wrote:
> > Hi, all
> >
> > What's the difference of the linux guest kernel parameter idle=<poll/mwait/halt/nomwait>, especially in performance?
> >
> > Taking the performance into account, which one is best?
> >
> > In my opinion, if the number of all VMs' vcpus is far more than that of pcpus, e.g. SPECVirt test, idle=halt is better for server's total throughput,
> > otherwise, e.g. in some CT scenario, the number of total vcpus is not greater than that of pcpus, idle=poll is better for server's total throughput,
> > because of less latency and VMEXIT.
>
> Makes sense overall.
>
> > linux-3.9 and above, idle=mwait is not recommended.
> >
> > Thanks,
> > Zhang Haoyu
>
> Does it actually have effect? I didn't think it would.
mwait instruction is never exposed to a guest. With idle=mwait is will
likely fall back to halt silently.
--
Gleb.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2013-11-21 7:22 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-11-21 3:45 [Qemu-devel] question about VM kernel parameter idle=<poll/mwait/halt/nomwait> Zhanghaoyu (A)
2013-11-21 7:01 ` Michael S. Tsirkin
2013-11-21 7:22 ` Gleb Natapov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).