Fred,
This is a problem we have observed as well. I'm sure you notice that
the CPU of your Windows VNC client goes through the roof as it attempts
to do things such as redraws. It is not a network issue at all, it is
in the Windows client itself.
I spent a lot of time working with the folks at Real VNC on this
problem, and you should see an dramatic improvement in their next
release.
- Jeff
On Mon, 2003-06-09 at 20:20, Fred Ma wrote:
> Hi,
>
> I tried using the Cadence CAD software using
> VNC. This is all in the same department, so
> so it should be at ethernet speeds. If I
> start up a vncserver on one sun box, and
> connect to it from another sun box, speed is
> fine. If I connect to a sun box from a PC
> using Exceed, speed is fine. If I connect
> to a sun box using VNC, but don't use Cadence,
> speed is fine.
>
> The only time speed is really slow is when I
> use Cadence via a VNC connection from PC to
> sun box. This is really strange, because if
> it was a bandwidth problem, it should exhibit
> itself when using VNC to connect from sun to
> sun. As well, it should show itself when
> using Exceed.
>
> Just for kicks, I Exceeded into a sun box and
> VNC'd from the sun box to another sun box. I
> launched Cadence in the 2nd sun box. Things
> were very sluggish. I then tried launching
> Cadence in the 2nd sun box, but as a normal
> X client to the X server on the 1st sunbox.
> Exceed still acts as the link from PC to the
> 1st sun box. It was somewhat slow, but not
> nearly as slow as the combining Exceed with
> VNC.
>
> I thought the bottleneck was conversion from
> Cadence's pseudocolor to the PC's true color.
> To test this, I creatd a 24-bit true color
> VNC server running on Sun. This server would
> emulate a PC. Within this server, I launched
> a viewer that connects to a 2nd remote
> pseudocolor VNC server. Speed of Cadence was
> fine, as long as I connect to the 24-bit VNC
> server from a sun box. Speed was terrible if
> I try from a PC.
>
> One thing I did notice when combining Exceed
> with VNC is this. Even though the local cursor
> (a tiny dot) is moved quickly to another location,
> the "other" cursor (the one which changes shape,
> presumably the "remote" cursor) lags behind and
> flickers very rapidly. As if the screen is
> undergoing many updates that prevent it from
> keeping up with the local cursor. It almost
> seems that the Cadence the application is not
> able to get information about the current cursor
> location until some lag time has passed.
>
> The only explanation I can think of for this
> disparity in speed is that Cadence uses high
> level draw/fill commands to the X server to
> update the window and buffer certain images.
> One example that supports this is the lightning
> fast response when switching between views that
> show only cell outlines versus cell details.
>
> Anyway, thanks if anyone can comment on this,
> or describe their own experience, including any
> experiments.
>
> Fred
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Etnus, makers of TotalView, The best
> thread debugger on the planet. Designed with thread debugging features
> you've never dreamed of, try TotalView 6 free at www.etnus.com.
> ___________________________________________________________
> TightVNC mailing list, [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/vnc-tight-list
_______________________________________________
VNC-List mailing list
[EMAIL PROTECTED]
To remove yourself from the list visit:
http://www.realvnc.com/mailman/listinfo/vnc-list
