System Freeze & Video Corruption

[fat-tony]

Hi!

I'm wondering if anyone can give me any suggestions. I've pretty much run out of ideas, but don't want to take this rig back unless anyone else has any ideas.

Hardware:
Asus A8N32-SLI Deluxe
Athlon64 X2 4400+
2Gb of generic PC3200 DDR (I'm going to have to dig up the invoice, I *think* I specified brand name)
eVGA GF7900GTX


Software:
XP Home (getting the free Campus Agreement Pro upgrade in a few days)
3DMark06
Memtest86

This all started out yesterday when the new system arrived at work here. First thing I did was install 3DMark06 to test out the machine. It ran smoothly for the first run, and then it went downhill from there. During GPU Test 4 "Deep Freeze", there begins to be video corruption, and then it gets very choppy and there are missing textures. When it bounces back into Windows, it is pretty much useless. The LCD keeps doing "Auto Image Adjust", and when the screen is actually ON, the windows are all jumbled up.

This brought an assumption of a messed up graphics card. I have tried updating all of the drivers, but this didn't help any.

I decided to do a RAM test, just to make sure. During the test, it through up an "Unexpected Interrupt":

Type: Gen_Prot
PC: 001000007
CS: 00000010
Eflag: 00010086
Code: 000024a4

Searches about this error haven't really turned up much. Does anyone have any suggestions? Right now I'm in the process of pulling RAM sticks to see if it makes any difference. This hardware is way newer than anything else around the lab though, so we can't really do much swapping around.

Thanks for any advice you guys might have!

 

[ChingChang]

Sounds like it could be the video card. What you described sounds like artifacts. Artifacts usually occur when a video card is overheating, or is overclocked too high. It can also occur from other things... and even simply reseating the video card can resolve the issue. GPU artifacts are messed up textures, and models "banding" across the screen", while video card RAM artifacts appear as white dots covering parts of textures.

I would reseat the video card and make sure there is good airflow to keep it cool. You can also try placing a household fan at the side of your case to see if helps. If it does, you might need to work on cooling. Or maybe there's a contact problem between the GPU and heatsink. If so you can reapply thermal compound or try an RMA.

Also do you have another video card to test with?

also if you wanted to test memory, you should try memtest86

 

[fat-tony]

Sounds like it could be the video card. What you described sounds like artifacts. Artifacts usually occur when a video card is overheating, or is overclocked too high. It can also occur from other things... and even simply reseating the video card can resolve the issue. GPU artifacts are messed up textures, and models "banding" across the screen", while video card RAM artifacts appear as white dots covering parts of textures.

I would reseat the video card and make sure there is good airflow to keep it cool. You can also try placing a household fan at the side of your case to see if helps. If it does, you might need to work on cooling. Or maybe there's a contact problem between the GPU and heatsink. If so you can reapply thermal compound or try an RMA.

Also do you have another video card to test with?

also if you wanted to test memory, you should try memtest86

Memtest86 is what I was using. I was very surprised that it got that General Protection Fault while running. That leads me to believe there may be something else at play here too.

Sadly, I don't have any other cards I can test with. This system is several years newer than anything else in the lab, so all of the others have AGP.

I put 8 hours of diagnostics into it yesterday, and it arrived in the lab the day before that. I'm just going to take it back and tell them not to deliver it until it works. In the end, the entire system cost us ~$2600CDN. There's NO REASON for it to come untested.

Edit: Oh yes, and I've already tried reseating the video card and the RAM. I just got an email back from eVGA tech support, and he suggests that it could be the +12V rail. I'll check that too before taking it back.

Hmmmm.... I wonder if we do have a big fan kicking around.

The strangest part about this is that it consistently fails at the *same* spot in the demo. Hopefully replacing the card will fix it...