NVIDIA 8800GTS Linux Issues

[nukem]

My girlfriend just built a new computer and we put Gentoo GNU/Linux on it. It has been running fine for the past few weeks but tonight when we went to look someone up we noticed that X had locked up so we restarted X(ctrl+alt+backspace) and saw this



I looked around and found in her dmesg this

NVRM: Xid (0001:00): 13, 0003 00000000 00005097 00000d00 3f800000 00000004
NVRM: Xid (0001:00): 30,  L1 -> L0

We posted on nvnews to see what nvidia says about this because she thinks its a driver error while I am thinking there might be something wrong with the card. What made me start to beleive this other then the lovely image we found was that she has an evga 680i motherboard. The chipset fan blows right onto the back of her card and I was thinking from the weeks of playing gaming and compiling it might have done something to the card. Does this error seem to be more hardware related, driver related, or only nvidia knows?

Thanks,

nuke

[nukem]

Well today we plugged it into my machine and it seems to be working fine so far(I get to try it for the night hehe). We are starting to think that it may be the power supply which is a Xion Supernova 600W. A review of it said that it has problems with power ripple suppression. Hopefully the motherboard wasn't damaged because of this.

[David]

seems like overheating to me. Overheating can get you artifacting and lockups. Also the power supply doesn't sound too reliable. I'd recommend something like fortron, sparkle, seasonic, antec, etc. The first two are usually less expensive.

[nukem]

It seems to have been the power supply. We had an old 420 Watt lying around and once we got all the converters for PCE-E and SATA we were able to test it and saw that all the errors went away. We just bought a few minutes before you posted a Corsair power supply. Hopefully no more problems will arise.

I doubt heating had anything to do with it the card ran hotter in my machine and her CPU and chipset also ran much cooler then mine.

Anyway thanks for your help!

[nukem]

Well it seems you were right. The problem came back after updating Gentoo. The chipset temps are very high (board 30C, MCP 51C). Not sure what to do I'm going to call evga tomorrow. The card still runs cool(according to nvidia-settings) so I'm saying the MCP is where the PCI-E controller is and is what is screwing it up.

Looking through newegg it seems that all the boards for Intel 775 have problems(I've read on everyone people saying they need to RMA 2 or 3 times). I don't remember this at all for boards for the AMD 939.

[nukem]

Well it seems that it was neither the power supply, a heating issue, or the card itself. I believe it is the nvidia drivers. According to this thread on nvnews.net Many people are having this issue with newer nvidia drivers. nvidia hasn't responded at all.

I'm wondering if anyone at Phoronix has had these issues. From the bug reports I saw it seems to appear mainly with nforce 680 chipsets and drivers in the 100 series.