Diagnostic suggestions, please

Yesterday my year and a half old game machine began to experience difficulties. The screen would scramble, with strange colors in bars and a pattern of white dots, sometimes over only half of the screen. I was playing Oblivion at the time and first thought it was a mod conflict, but it has also happened in Windows, so that isn't it. If I don't reset and reboot the machine eventually locks up completely. It's done this several times now. This machine is a quad core processor running on an Asus P5N-D motherboard with 2 gigs of Kingston ram and an Nvidia 8800GT video card, powered by an NZXT PP-600 power supply. Following the crash I went to BIOS and looked at the hardware monitor, checking temperatures and voltages. All of the voltages were within a few thousandths of nominal, the mobo was at room temperature, more or less, and the processor at 48 C, a little over 100 F. I wouldn't think that should be a problem. The video card felt pretty hot, so I removed and disassembled it. I cleaned the fan & heat sink, re-greased it and put it back together and back in. It did seem to run cooler, but I still crashed within a few minutes.
My first thought was that the video card is dying, even after cleaning. However, that first crash while playing Oblivion corrupted the savegame I made at about the time it crashed, and I wouldn't think a video card would do that. I might not be able to see what was happening, but the save process should still have worked.
Any suggestions as to where to go from here?

 

Further update. I read about OCCT in another thread on this forum. I downloaded it (from the authors website, the Major Geeks links don't work) and ran it. After about 6-8 minutes of CPU test my CPU temp had stabilized at about 58C and stayed there. I shut that off and started the power supply test. I got an immediate screen scramble and lock up. Reset & reload. Tried the GPU OCCT test. Another immediate scramble & lock up. Let things sit a few minutes and tried to re-start. Now the screen is scrambled on start up, so I just shut it off.
Question: can I interpret this as a death in my video card family or is it still possible that this is a ram, or other component, issue?

 

Ouch man, the OCCT was probably mentioned by me. I'd agree with David as to it's the GFX card. I'm surprised that it blew after some 3 or 4 years but crap happens.:cry

You can test your memory using Memtest86 once you get a replacement card, just to be thorough. Your CPU temperature is excellent and you have a decent power supply. Be careful with OCCT, just run it for 24hrs on new stuff when needed so you can RMA it if needed. That's what I do whenever I build a new machine. Good luck.

 

I can't remember the last time I had a video card die, other than right out of the box, either. As you say, however, spit happens. Since I've been thinking of going to a somewhat newer card anyway, if something had to go, the GPU would be my first choice. So, off to buy a new card. Thanks for your advice and support.
I just noticed another reply in this thread suggesting a testing procedure. Unfortunately, my only other computer is an antique without a PCI Express slot, and I don't have access to another card. Yet. So be it. Thanks again, gentlemen.

 

Let us know what happens, please.

 

The GPU in question was an Nvidia 8800GT that came with the machine. I replaced it with an Nvidia GTS 450 from the local Fry's. That is about as strong a card as I can use with out also upgrading the power supply, since my 12 volt rail is rated at 22 amps and everything from 460 on up wants 24 amps or more. Stability and goodness have returned.
Thanks again for all the help. I don't visit this site as often as I used to since I can't afford to build, or even rebuild anymore, but sure appreciate the assistance.

 

Good choice for the video card but I'd not depend on an iffy power supply being run near it's max, especially since it's old as they lose some %10 per year.