Hi, we have ATI FireGL and Mobility cards in our machines, which hasn’t got the new drivers yet, and it usually takes a while…
I just wonder if the color swich bug and/or some other things has been fixed in the newest ATI Catalyst driver, which suppose to have fixed at least some OpenGL stuff.
It has OpenGL 2.0 support included as well, which i guess is good?
Cheers,
Mikael Grev
Unfortunately there is no hardware support on any Ati chipset for non-power-of-2 texture sizes.
From the release notes:
“The CATALYST™ Software Suite introduces OpenGL Version 2.0 support. This version of OpenGL will be in line with the latest GL ARB approved specifications. New to this specification are GLSL, non-power-of-two-textures, separate stencil, multiple render targets, and point sprites”
So, they seem to have fixed that power of two thing, at least on the software side… 
Cheers,
A friend of my has ATIRadeon9800 card and it sux on dualview desktop. All OpenGL (ogre3d, delphiGL, jogl, lwjgl) apps run fine on primary desktop.
But soon as I move window to the secondary desktop, it just freezes. Actually app is running fine, but only that ATI dispayes the freezed frame was visible first time.
I then uploaded the latest WQ certified driver package and installed it. Well, now OpenGL run on both desktops. But problem was both 1st and 2nd desktops run it like 5 fps.
DirectX work better, but applications that do not give ability to select the device context may run slower on 2nd desktop. Ogre3d demos have a startup dialog where I can select Radeon1 or Radeon2 devices. It fixes fps on 2nd desktop, but then I cannot move it to 1st desktop 'cause it drops to 10 fps.
To be honest I am a bit disappointed about the Java2D-opengl implementation.
I thought it would solve the problems java has on X11 with a bit weak java2d performance compared to Win32.
Since I have a nvidia card which are known to have the best available drivers for Linux - I was really sad when the first version didn’t even work at al.
I wrote the nvidia devs and the second version was useable, but most stuff ist still faster when doing rendering antialiased - cause then its rendered in software 
I do not want to say that its SUN’s fault or NVidias fault, I think they are both repsonsible for this.
The ogl-pipeline is not major enough, I encountered some bugs which are completly independent of the underlaying hardware/driver combination and nvidia whith their drivers only optimized for gaming.
just my 2 cents…
lg Clemens
[quote]I just wonder if the color swich bug and/or some other things has been fixed in the newest ATI Catalyst driver, which suppose to have fixed at least some OpenGL stuff.
[/quote]
Hi Mikael,
I installed the latest (5.3) Catalyst drivers on a couple of our machines (Radeon 9500 Pro, 9800 Pro), and no, the driver bugs affecting the OGL pipeline are still not fixed. I finally found the time to file 3 bugs against ATI, and I know that they’re starting to look into them. Hopefully it won’t be too much longer until they’re fixed. The fixes should be in shared code, so the same problems on Linux should be fixed as well.
In the meantime, you might want to track these two Java bugs, which reference the relevant ATI driver bugs…
Gradient/Color swap problems:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=5033205
Crash on exit bug:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=5066318
We just finished up a big revamp of the OGL pipeline, which means increased stability and performance in Mustang. I’ll have a new blog entry up on Monday with further details (and good news about J2DBench too).
Thanks,
Chris
They’re still unusable for GLSL, at least for what we’re doing.
Some examples:
Bugs 2 & 3 are there since their first implementation.
Oh, with the 5.3s most shaders crash the VM when linked. That’s a new “feature”. I guess we’ll be able to release Marathon in a couple of years…
Hi Chris,
Thanks for the info.
Btw,
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6219284
has some really nice microbenchmark figures… Well done!
You get things done Chris
Cheers,
Mikael