AFAIK, right now both the Xen hypervisor and the guests run with a HZ value of 100. Is there any fundamental problem running them at 1000 instead? (well I am trying it now, so I''ll know if it even runs at all, but I just wanted to know if there were any deeper issues) -- Diwaker Gupta http://resolute.ucsd.edu/diwaker ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
Xen is run with HZ=100 but this should only be used for time keeping. Most of the timer functions are actually hung off a reprogrammed APIC timer. For guests we ideally want them to tell xen their next timeout value (which they should know) and then we''d program a timeout value for them . IIRC a guest gets periodic timer interrupts when it actually has the CPU and we fake out he missed ticks when scheduling a VM. As the Linux/370 people have pointed out a while ago (a few years) sending a periodic timer to quiescent VMs is not scalable as the system will spent an awful lot of time just processing pointless timer interrupts for these VMs. For xenLinux 2.4 we therefore used the IBM patch (to the timer code) to program the APIC timer (via Xen). I don''t remember what the current 2.6 story is. I believe there is an effort to push these patches in main stream Linux. With all this in mind, why would you want to increase the HZ value? Thanks rolf> -----Original Message----- > From: xen-devel-admin@lists.sourceforge.net [mailto:xen-devel- > admin@lists.sourceforge.net] On Behalf Of Diwaker Gupta > Sent: 10 March 2005 19:25 > To: xen-devel@lists.sourceforge.net > Subject: [Xen-devel] HZ=1000 with Xen? > > AFAIK, right now both the Xen hypervisor and the guests run with a HZ > value of 100. Is there any fundamental problem running them at 1000 > instead? (well I am trying it now, so I''ll know if it even runs at > all, but I just wanted to know if there were any deeper issues) > > -- > Diwaker Gupta > http://resolute.ucsd.edu/diwaker > > > ------------------------------------------------------- > SF email is sponsored by - The IT Product Guide > Read honest & candid reviews on hundreds of IT Products from realusers.> Discover which products truly live up to the hype. Start reading now. > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/xen-devel------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_ide95&alloc_id396&op=click _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
Hi, On Thu, Mar 10, 2005 at 11:24:49AM -0800, Diwaker Gupta wrote:> AFAIK, right now both the Xen hypervisor and the guests run with a HZ > value of 100. Is there any fundamental problem running them at 1000 > instead? (well I am trying it now, so I''ll know if it even runs at > all, but I just wanted to know if there were any deeper issues)1000 timer interrupts per second sent to every domain would be detrimental to performance if you think of many domains. Regards, -- Kurt Garloff <kurt@garloff.de> [Koeln, DE] Physics:Plasma modeling <garloff@plasimo.phys.tue.nl> [TU Eindhoven, NL] Linux: SUSE Labs (Director) <garloff@suse.de> [Novell Inc]