Video Card Failing - How do I diagnose?

Over the last few days I have been getting random and very intermittent BSOD's for the first time with Win7 x64.

Each BSOD during the memory dump before rebooting was making a reference that it couldn't start up my video driver or something in that nature. I have a nVidia evga 9800 GX2 with 190.37 drivers.

After a BSOD this morning, same reference as above, windows rebooted but resolution was 640x480, a quick look under device manager showed both GPU's on my 9800 GX2 had been stopped due to a problem with exclamation marks beside them but oddly I still had a display.

After the next reboot, I couldn't get into windows anymore. Not windows 7, or XP, or leopard (I am Triple Booting) I CAN however get in via safemode but my troubleshooting is not revealing much cause in safe mode, those drivers are naturally disabled.

The event viewer under administrative tools, showed a kernel-power reference during my last couple BSOD's which I believe is just me hitting the reset button when the computer was freezing up. There are no other critical references.

My initial thoughts is that something in the video card was slowly beginning to go, hence the casual BSOD's over the past couple days and now that's it. BUT the odd thing is that ALL the LED's on my video card are showing correctly, ie, if there was some kind of internal problem, there are a couple of them that should be showing red or yellow, but they are not.

Anyone have anyway of diagnosing that it is indeed my video card that is failing? Could it be the motherboard? Also, in safe mode in Win7, I tried reinstalling the 190.37 drivers but no changes.

Thoughts? Questions?

Thanks a bunch.

PS - I have tried startup (repair) and I have tried restoring to an earlier point. No changes. I am still in safe mode.

After shutting it down and letting it cool off for about an hour I was able to get back into windows 7 with a normal boot BUT with the issues I described above, low resolution, device manager shows 9800 GX2 has been turned off, the highest res I can get is 1280x1024 and my monitor is not recognized as a BenQ anymore, it just says default.

I'm having a similar problem as you, but I think that Mine may be a little more advanced -- my card isn't showing up at ALL. My advice, test your cards on another machine, and move forward extremely gingerly.

Thanks for the reply. I don't have another machine handy to give it a try but I have tried switching PCI-E slots (my mobo has 4) and still the same result. Also as previously mentioned, I have tried booting into all 3 of my OS's with no success so I think this rules out the chance of a driver or software issue.

Thing is, it could also be my motherboard, or PSU, but I don't think it's likely considering everything else works normally and windows will boot up in Safe mode or boot up as I previously described.

Like I said, all the LED's on the card itself are normal and there are 6 of them I believe! This is what is kind of throwing me for a loop.

You'd be surprise what a bad mobo or ram could do to a system. The only way you're going to be sure that it's the card itself is to test it on another machine. Good luck mate.

TheMumm said:

You'd be surprise what a bad mobo or ram could do to a system. The only way you're going to be sure that it's the card itself is to test it on another machine. Good luck mate.

Actually I wouldn't I have experienced and troubleshooted tons of tricky and incognito ram related issues in my past.

If it was my ram I would still be having issues running my computer in safe mode but I am currently not.

Have you set it back to the normal clock speeds? If you can do this in safe mode. Never mind I didn't look at your pics in link #2 GPU-z reports 600/1000. OH you stated your monitor is now reconized as VGA PNP? Which can cause resolution problems if the driver for your monitor (LCD) isin't installed. You also might have borked it with your overclock, or heat related but I've found out when it comes to video cards when they go they just go. Hardly ever any rime or reason except overcloked and or over heated. Fortunetly, if you registered it in the 30 day period after purchase they will replace it but in the mean time you'll need a card or pay the cross shipping fee. I'd try fixing the monitor driver situation first.

Ok few things.

I always keep it stock speed and only OC it when needed as you say you saw in the pic but in the pic, not sure if you noticed, it showed the 'default' clock as 600 and the actual at that time as zero! (compare with the new one below but read first).

The monitor was not recognized only AFTER I had BSOD and GPU issue.

Here's the strange part, as stated earlier, I have tried booting into various OS's and stated I could only get in via SAFE mode with it SOMETIMES booting up normally but with no video drivers or good resolutions functioning.

I booted back to Win7 safemode and decided to completely uninstall the drivers, and then run NFR (nasty file remover), cleaned out all traces of Nvidia stuff and reinstalled 190.37 64bit.

I am now rebooted and back into windows 7 with my video card functioning it would appear normally?!?! (see attached) with my BenQ monitor showing up as well under the resolution window.

I did the same under WinXP, and progressed as well but didn't quite get the same results, device manager under XP reports 1 of the 2 GPU's not functioning but oddly Windows 7 does not.

Before all this I had finished running memtest86+ for hours and it passed 3 loops so my ram is definitely not an issue just to rule this one out.

Also, I had already called evga and gotten an RMA with a free upgrade to the GTX 280! Not sure how exactly the 280 fairs against the 9800 GX2 so I don't know if I should keep the card if it's functioning, or just do the RMA.

Bottom line, is this an intermittent problem that is just unstable and ready to give eventually with one of the card's GPU's or is it possible that somehow the drivers got corrupted and needed a reinstall. But if the latter is the case, how is it possible that it was required for both Win7 and XP???

If it were my card I would stress it by playing a game on high settings and see if it starts to artifact or act up. this would give you a good idea if one of the gpu's is failing or run some benchmarks on it, get it hot. If it is stable then keep it. or do some research on a comparison to the two cards and decide that question. The 280 is a totally different core and from what I have read on Toms it out preforms the card you have. So it's up to you. If the problem occurs again then try and zip a minidump file and post it. I switched from 2 9800 gt's to a 260 superclocked and it's awesome. Fabe

Before suspecting one of the GPUs, I'd look at the cards' capacitors. Ensure the tops are flat and not rounded. Make sure there's no brownish or discolored substance leaking from any of them, near the base and top. If so, that's the electrolyte leaking out. Not good.

The caps are partially responsible for regulating power. If one or more of them go, Windows can exhibit all sorts of behavior.

The plus side to this is that caps are easily replaced with basic soldering. Nothing difficult, if this is indeed the case.

If you do replace any capacitors, purchase ones with the same capacitance of course but also make sure they are the low esr type. Your card will enjoy the efficiency. And when installing, make sure the colored part of the capacitor lines up with the colored part on the board. That's for correct polarity.