Re: [Xen-users] iozone on HVM guest

["Rami Rosen" <roszenrami@xxxxxxxxx>]

Hi,
You can Look at diagrams 5 and 6 and the corresponding sections in:

"High-Performance Hypervisor Architectures: Virtualization in HPC
Systems" paper

>From the 1st Workshop on System-level Virtualization for High
Performance Computing
(HPCVirt 2007)
held in conjunction with EuroSys 2007,

see : http://www.csm.ornl.gov/srt/hpcvirt07/papers/paper10.pdf


Regards,
Rami Rosen


On Dec 5, 2007 12:28 PM, Pasi Kärkkäinen <pasik@xxxxxx> wrote:
> On Wed, Dec 05, 2007 at 09:48:52AM +0200, Ofek Doron [Ofek BIZ] wrote:
> > Hi all,
> >
> >
> > I am looking for information / results of iozone in a linux HVM based
> > guest.
> > I tried google but didn't find .
> > Does anyone have a link (or file) to such kind of information ?
> >
> > also, looking for objective information about I/O performance in a
> > paravirtual guest vs. HVM based guest.
> >
>
> Some results with Oracle OLTP workload:
>
> http://www.redhatmagazine.com/2007/11/20/red-hat-enterprise-linux-51-utilizes-nested-paging-on-amd-barcelona-processor-to-improve-performance-of-virtualized-guests/
>
> Some other workloads might and will be different.
>
> PV guest is 91% of the baremetal speed, and HVM guest _with_ PV drivers
> (important!) is 75% of the speed of baremetal.
>
> HVM guest performance without PV drivers will be sucky, so make sure you're
> using them.
>
> -- Pasi
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users
>

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users