<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Performance isn't optimal right now as there's a lot of excessive
buffer copying and rescheduling. The current<br>
code is just a proof of concept. When things calm down here, I'll try
to find some time to rewrite it to use command<br>
and display buffers in guest VRAM. <br>
<br>
Opengl extentions are pretty much unused right now. Only the basic
functions are implemented. We have to look at the<br>
various opensource solutions to see which one fits best.<br>
<br>
Ticket 475 contains all the information you need to get the OpenGL ICD
up and running. What is missing?<br>
<br>
Ranando King wrote:
<blockquote cite="mid:200708280219.11212.rwking@usa.net" type="cite">
<pre wrap="">I noticed that when I looked throught the code. That stuff is a nightmare to
trace through!!! But anyway, I 've been using a demo/benchmark program I
found at <a class="moz-txt-link-freetext" href="http://www.glexcess.com/defaultn.asp">http://www.glexcess.com/defaultn.asp</a> to test how well the
implementation is working. As of tonight's SVN, I haven't been able to
reproduce the problem I mentioned.
At the same time, I'm getting very low frame rates. By low, I mean less than
10 in many cases on this application when I run it full screen at
1280x1024x32. I did notice something when I ran it at 640x480x32. It seemed
to run relatively smooth for short periods, then slow down. At the same time,
the fps meter would show high frame rates during the slow period and low
rates during the smooth period. I guess I still don't have the OpenGL
extensions configured properly. I followed the instructions from bug report
475 in my attempt to configure.
So if someone will please list the instructions for configuring the OpenGL
additions after building them, I would greatly appreciate it.
R.
On Monday 27 August 2007 06:40, you wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Hi Ranando,
I haven't seen that behaviour here yet. Typically applications use a
blocking flush to empty the command queue
and that blocks the thread in the guest OS as well.
Ranando King wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Never mind. I commented out some restricting lines in VMMDevInterface.cpp
and got it going. But I'm noticing something. I'm using the GLXS
benchmark to test performance. While the benchmark is reporting something
between 20 & 115 fps for complex scenes, I cansee that the display is not
keeping up. The display appears to be giving me something more akin to
3-5 fps. When I look in the log file for the running session, I see a lot
of this:
</pre>
</blockquote>
</blockquote>
<pre wrap=""><!---->
_______________________________________________
vbox-dev mailing list
<a class="moz-txt-link-abbreviated" href="mailto:vbox-dev@virtualbox.org">vbox-dev@virtualbox.org</a>
<a class="moz-txt-link-freetext" href="http://vbox.innotek.de/mailman/listinfo/vbox-dev">http://vbox.innotek.de/mailman/listinfo/vbox-dev</a>
</pre>
</blockquote>
<br>
<br>
<pre class="moz-signature" cols="72">--
Kind regards / mit freundlichen Gruessen / Met vriendelijke groet
Sander van Leeuwen
innoTek GmbH
<a class="moz-txt-link-freetext" href="http://www.innotek.de">http://www.innotek.de</a>
Germany </pre>
</body>
</html>