Re: [PATCH 0/9][RFC] KVM virtio_net performance

[Anthony Liguori <anthony@codemonkey.ws>]

Mark McLoughlin wrote:
> Hey,
> 	One all all-important thing I forgot to include was a comparison with
> lguest :-)
>   

Hey Mark,

This patch set is really great!  I guess the hard part now is deciding 
what all we want to apply.  Do you have a suggestion of which patches 
you think are worth applying?

BTW, do you have native and guest loopback numbers to compare where we 
stand with native?

>   netperf, 10x20s runs (Gb/s)  |       guest->host          |       host->guest
>   -----------------------------+----------------------------+---------------------------
>   KVM                          | 4.230/ 4.619/ 4.780/ 0.155 | 8.140/ 8.578/ 8.770/ 0.162
>   lguest                       | 5.700/ 5.926/ 6.150/ 0.132 | 8.680/ 9.073/ 9.320/ 0.205
>
>   ping -f -c 100000 (ms)       |       guest->host          |       host->guest
>   -----------------------------+----------------------------+---------------------------
>   KVM                          | 0.199/ 0.326/ 7.698/ 0.744 | 0.199/ 0.245/ 0.402/ 0.022
>   lguest                       | 0.022/ 0.055/ 0.467/ 0.019 | 0.019/ 0.046/89.249/ 0.448
>
>
> 	So, puppies gets you an extra 1.3Gb/s guest->host, .5Gb/s host->guest
> and much better latency.
>   

I'm surprised lguest gets an extra 1.3gb guest->host.  Any idea of where 
we're loosing it?

> 	Actually, I guess the main reason for the latency difference is that
> when lguest gets notified on the tx ring, it immediately sends whatever
> is available and then starts a timer. KVM doesn't send anything until
> it's tx timer fires or the ring is full.
>   

Yes, we should definitely do that.  It will make ping appear to be a lot 
faster than it really is :-)

Regards,

Anthony Liguori

> Cheers,
> Mark.
>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>