[Spice-devel] Unfair comparisons with RDP

[Andrew Cathrow]

----- Original Message -----
> From: "John A. Sullivan III" <jsullivan at opensourcedevel.com>
> To: "Alon Levy" <alevy at redhat.com>
> Cc: spice-devel at lists.freedesktop.org
> Sent: Wednesday, June 29, 2011 6:14:16 PM
> Subject: Re: [Spice-devel] Unfair comparisons with RDP
> On Sat, 2011-06-25 at 11:40 -0400, John A. Sullivan III wrote:
> > On Sat, 2011-06-25 at 15:34 +0200, Alon Levy wrote:
> > > On Fri, Jun 24, 2011 at 09:36:26PM -0400, John A. Sullivan III
> > > wrote:
> > > > Hello, all. Another question to be answered only if there is
> > > > time. As
> > > > we put SPICE head to head with the TSPlus implementation of RDP,
> > > > SPICE
> > > > seems to be coming up a bit short and I expect it is somewhat
> > > > unfairly
> > > > so. It appears that RDP starts painting the screen sooner that
> > > > SPICE
> > > > does and so it appears more responsive even if it takes longer
> > > > to finish
> > > > painting the page than SPICE does.
> > > >
> > > > Why is that? More importantly, is there a way to tweak SPICE to
> > > > mitigate
> > > > this effect or is there something in the works to make the
> > > > initial delay
> > > > less pronounced? Thanks - John
> > >
> > > What exactly is the test? what do you mean starting / stopping,
> > > does the
> > > time include the launch of spicec or spicy and the connection?
> > >
> > > If it includes the connection time perhaps this is a result of the
> > > insane
> > > bandwidth calculation we use, sending a very large packet of zeros
> > > at
> > > connection initiation.
> > <snip>
> > Sorry for the lack of clarity. I'm not referring to start up time.
> > That IS noticeably slower but I think the bandwidth detection is a
> > critical feature and differentiator for SPICE.
> >
> > I'm referring to actual usage of the desktop. All actions from
> > scrolling a window to clicking a slider to opening a hyperlink, even
> > typing very quickly seem slower to start in SPICE than in RDP. The
> > difference is only a fraction of a second but it is humanly
> > perceptible
> > and gives the impression of laggard performance.
> >
> > On large graphic refreshes of the screen, it seems clear that SPICE
> > finishes painting the screen faster than RDP even if it appears to
> > start
> > later - it comes from behind like the Boston Red Sox (sorry for the
> > American sports humor!) but appears to the user who is more
> > concerned
> > about when it starts to be slower. When changing only small parts of
> > the screen such as typing, there isn't the opportunity to "come from
> > behind" and thus SPICE is being rated as slower than RDP, i.e, for
> > everyday usage, not start up. Thanks - John
> <snip>
> Hello, all. I've been using both RDP via TSPlus and SPICE for over a
> week now and the practical world results at least in my mode of
> operation are becoming clear. SPICE does handle major screen refreshes
> better, e.g., monstrous graphics or continuously pasting a full line
> of
> text in a full screen notepad. Of course, SPICE is a clear winner when
> it comes to video though still not practically usable on low bandwidth
> links.
> 
> However, RDP is trouncing SPICE in the more common day to day tasks
> involving small screen updates. For example, every time I open or
> switch
> my screen to LibreOffice, the tool bar icons seems to pain one at a
> time
> in SPICE where as they appear all at once in TSPlus. Document
> scrolling
> is more immediate and smoother. Surprisingly, when I open the PuTTY
> dialog in SPICE, it paints in sections whereas TSPlus appears all at
> once.

You'll really need to post some real details here - from versions of qemu+spice server that you're using along with the qemu command line syntax, host OS, hardware details through to what's running in the guest, driver versions, etc.

I've never seen RDP trounce spice, but if it does, then we need some scientific information on the environment so we can diagnose and assist.




> 
> Certainly not meant as a competition but merely as providing feedback
> from practical usage. We remain very enthusiastic about SPICE but
> wonder what we can do to close the performance gap on the more common
> tasks (in our usage) versus the more unusual tasks.
> 
> By the way, I did set tcp_low_latency=1 on my client to see if it made
> a
> difference. It subjectively appeared to but not enough to close the
> gap. I should also mention that my entire test environment is Linux
> clients. Is the Windows SPICE client noticeably faster? We are in the
> process of lining up some Windows client testers. Thanks - John
> 
> _______________________________________________
> Spice-devel mailing list
> Spice-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/spice-devel