Hugo Mills wrote:
> On Thu, Mar 13, 2008 at 09:24:02PM +0000, Simon Capstick wrote:
>> Hugo Mills wrote:
>>> On Thu, Mar 13, 2008 at 08:22:16PM +0000, Simon Capstick wrote:
>>>> It looks like my VIA C7 just doesn't have the oomph for kqemu to run
>>>> Windows XP fast enough to be useful. A faster Intel/AMD processor with
>>>> virtualisation extensions looks necessary for this. I guess I will be
>>>> powering up a noisy XP box occasionally instead.
>>> You can run qemu-based systems (qemu,kqemu,kvm) on a remote server
>>> and connect to them over VNC, if that helps.
>>>
>>>> Am I right in presuming I can't run KQEMU or KVM on a (noisy but remote)
>>>> server running the Xen hypervisor?
>>> kqemu *may* work, but I wouldn't guarantee it. KVM almost certainly
>>> won't work.
>> That's what I thought, thanks for confirming this. For the time being I
>> have managed to speed up KQEMU a bit by disabling ACPI [1] in the
>> Windows XP VM. The trouble is the qcow2 image is on a LUKS encrypted
>> external disk thus crippling my performance further.
>
> This may be of interest to you:
>
> http://article.gmane.org/gmane.comp.emulators.kvm.devel/14393
>
> That mail says it's not ready yet, but worth keeping an eye on.
>
> I got a rough 2x performance increase on a kernel compile using
> virtio for a Linux guest -- you'll find my mail on how I did it in
> last months' HantsLUG list archives, and here[1].
>
> Hugo.
>
> [1] http://www.carfax.org.uk/docs/qemu-virtio/
>
>
All very interesting, thanks Hugo. Since I'm having problems even
getting tun/tap working with kqemu, probably due to my weird desktop
set-up, I'll postpone further experimentation until I dig out a
dedicated and faster box to happily play on without risk.
Thanks,
Simon
