Chequerboard pattern distortion when using DRI
Hi all,
I've managed to get DRI working on my machine......sort of. The 3d acceleration works alright, but laid over it I get this wierd chequerboard-like pattern. In Tuxracer the distortion manages to obscure half the screen. To illustrate the problem, I have put a snapshot on the web: http://homepage.ntlworld.com/richard-wild/snapshot1.png I have two questions. What on earth is wrong? What should I do to fix it? TIA. |
What hardware do you have, what distro and what drivers are you working with????
It could be one of 3 things: Bad hardware, wrong driver, bad setup of XFree86. I'm leaning towards the second two. We need more info to analyze your problem. |
Video card is a Sapphire ATI Radeon 9800SE. Motherboard is an Asus A7N8X-E Deluxe, which uses the NVidia NForce2 chipset. I am using the ATI proprietary driver, version 3.9.0 which is very new (posted a few days ago I think).
Distro is Mandrake 10.0, running a 2.6.3 kernel I compiled myself from the sources included with the distro. I compiled my own in order to have agpgart and the nforce2 driver in the kernel. One thing I will certainly try this evening is to obtain an older version of the ATI driver and see how well that works. Thanks for your interest! |
Bad news. I tried installing the older version of the ATI driver, version 3.7.6 (previously I was using the new 3.9.0 driver). I still get the same problem.
I do hope my graphics card isn't fubar. That would royally suck. This weekend I will take my PC to my mum's house and try my brother's graphics card in my machine (and my card in his pc too). See what happens that way. |
Bump
|
I wouldn't have compiled AGPgart and the video driver into the kernel. I would have just installed them as kernel modules. It's easier to swap them out that way with new drivers.
Try going back to a stock 2.6 kernel and using the ATI drivers as well as AGPgart as a module. |
I tried it with agpgart and the nforce2 driver as modules too, same problem. I will go right back to the 3.2.8 driver and try that before I start despairing.
The only place I could find any reference to this problem [other than another brief mention in another thread here where it was only mentioned in passing] is on a forum that was discussing the soft-modding of the 9800SE card into a 9800 or 9800Pro. Apparently enabling the 4 deactivated rendering pipelines to make it an 8-pipeline card can cause artifacts because of undocumented defects with the quad-pipe cards. But I am trying to run mine as God intended! |
Bump
|
Update time. At my brother's suggestion, we installed WinXP on my machine (eeew) and found that Enemy Territory looks fine. So there is nothing wrong with my hardware. I am now hastily reinstalling Mandrake and will go and take five baths.
So something in my X setup is causing these artifacts. |
We never did understand what was causing the artifacts. My brother very graciously agreed to swap his Radeon 9600Pro, which does work on my linux machine, for my Radeon 9800SE, which works fine under Windows. I think he got shortchanged there, so I am very grateful to him. The problem could not be said to be sorted though - he has a Mandrake 9.1 dual-boot on his Windows machine, and he gets the same problem with that card when he boots into Linux.
I would be very interested to know it this is a general problem with the Radeon 9800SE, or whether it is just that particular card. |
I have the same problem.
I have exactly the same problem as you, Radeon 9800 SE has graphics artifacts in linux, but not in Windows. I also have an Nforce2 motherboard, but I don't know if that is related to the problem. Until now, I haven't been able to find any mention of this problem on the net. I just registered on these forums hoping that we can find a solution, but I suspect it might be something only ATI can fix. I can also verify that the artifacts I get in linux are the same as when I try the soft modded drivers in Windows. I knew I should have got an Nvidia card... :cry:
|
Re: I have the same problem.
Quote:
It seems possible, then, that the Linux drivers are enabling the deactivated pipelines. This ought to be something that ATI can fix, if they care. This is probably worth a friendly email to them. Does anybody think my logic is bogus here? |
All times are GMT -5. The time now is 03:28 PM. |