[vfio-users] Physical vs virtual machine gaming performance comparison benchmarks

[Blank Field]

So you wrote a web control panel for libvirt, docker and some NAS, right?
Can you try providing VDI via Spice or VNC on gigabit LAN?
On Sep 6, 2015 2:38 PM, "Jon Panozzo" <jonp at lime-technology.com> wrote:

>
> On Sep 6, 2015 5:54 AM, "Blank Field" <ihatethisfield at gmail.com> wrote:
> >
> > Have you used cpu pinning?
>
> Yes.  Our OS automatically pins CPUs for VMs.
>
> > Do you attach a screen to your "NAS" box or you connect remotely? If
> using remote connection, what is it, VNC, Spice or some other?
>
> Yes, local graphics output.  No remote connection used, although I also
> leverage Steam in home streaming from other devices for gaming as well.
> Our primary use case is to allow folks to add NAS capability to a
> performance gaming rig.  We also have Docker support.  This allows 3 tiers
> of functionality: network attached storage, virtual applications, and
> virtual machines.
>
> > Did you enable HugePages?
>
> No, we don't have hugepages enabled.  We tested hugepages before, but
> honestly didn't see any gains.  We do have transparent pages enabled on the
> host, but our guests do not make use of the <hugepages> XML tag (we use
> libvirt).
>
> > How much memory the VM had allocated?
>
> The blog post covers this:
>
> "...other thing worth noting here was that for our bare metal test, all
> 16GB of RAM was accessible to the OS, but with our VM tests, we only
> allotted 8 of the 16GB of RAM."
>
> >
> > On Sep 6, 2015 12:50 PM, "Jon Panozzo" <jonp at lime-technology.com> wrote:
> >>
> >> Hello vfio users!
> >>
> >> A common question folks seem to have is what sort of performance they
> can expect out of a VM for gaming as compared to that of a physical
> machine. My company makes a linux-based OS that features the capabilities
> of a NAS and a virtualization host using KVM, which has been tuned for
> maximum compatibility with vfio GPU assignment.
> >>
> >> In a recent blog post, I've shown benchmark testing results in
> comparing a physical machine to a virtual one using the same hardware and
> with varying VM configurations. I even put together a video clip showing a
> side by side comparison of the benchmarks as they ran in real time,
> allowing you to visually see the difference in each test. Thought this
> might be of interest to some of you, do here's the link:
> >>
> >> http://lime-technology.com/gaming-on-a-nas-you-better-believe-it/
> >>
> >> Tl;Dr:  I achieved up to 98% of bare metal.
> >>
> >> Regards,
> >>
> >> Jonathan Panozzo
> >> Lime Technology, Inc.
> >>
> >>
> >> _______________________________________________
> >> vfio-users mailing list
> >> vfio-users at redhat.com
> >> https://www.redhat.com/mailman/listinfo/vfio-users
> >>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/vfio-users/attachments/20150906/7349f6a2/attachment.htm>