Help, I think my 7950Gt is dieing

I believe that my card is on its last leg. Its been running at the speeds in my sig since last christmas, and just recently I have been getting system lockups while gaming. Ive already ruled out the PSU, CPU is Prime stable, its not the memory. Could it really be that my card has finally said it doesnt want to run at these speeds anymore, and is locking up my system.

Assuming you're running an NT-based OS like XP... open Event Viewer (Start | Run | Type eventvwr.msc in the Run box) and right click--> "Properties" the latest critical errors you've received (critical errors are in Red) around the time of the "system lock-ups". Copy and paste (use the "Copy to Clipboard" button under the arrow buttons in the "Event Properties" window) a couple of the latest critical error message(s) as they appear under the "Description" field, and post them in this thread.

After you hit the "Copy to Clipboard" button in the "Event Properties" dialog, the contents in the Description field will be copied to the ClipBook Viewer (ClipBoard). You can then just paste the contents into your post, as there is no Edit-->Select All-->Copy function in ClipBook. In addition you could type clipbrd in the Run box, which will open the ClipBook Viewer containing the copied text from the Event Properties dialog... then perform a File-->Save As, and save the file with a *.txt extension. You could then perform a Edit-->Select All, then do a copy/paste of the text in your new post... so you have a couple different options.

Also, open the System Information tool by typing msinfo32 in the Run box | Expand "Components" | Highlight "Problem Devices", and look for any problems that might be listed there.

I tried looking through the Even viewer and I didnt see anything in Red for errors, I only saw a couple of yellow flagged ones but they were internet related. Also, nothing came up under system information in the problem devices area either.

Try blowing out all the dust. Dust could be the problem.

There isnt any dust eihter lol, its all clean.

When you say "CPU is Prime stable", for how long did you run it, and are you using Orthos (or two instances of Prime95)? I would say you should run Orthos torture test for at least 8 hours before you conclude that CPU/memory are okay. After doing that, I would recommend is looping 3DMark2001 for several hours. If the system crashes/reboots while looping, the video card is probably bad. Let us know what you learn.

Well, the longest Ive ever gotten to run Prime is about 6 hours before someone came in my room and turned off the damn PC. And yes, I run two instances of Prime to test it for stability. I did run a couple loops of 3DMark06, and it failed just at the end stating it was a device error.

Have you tryed new drivers?

KillrBuckeye said:

When you say "CPU is Prime stable", for how long did you run it, and are you using Orthos (or two instances of Prime95)? I would say you should run Orthos torture test for at least 8 hours before you conclude that CPU/memory are okay. After doing that, I would recommend is looping 3DMark2001 for several hours. If the system crashes/reboots while looping, the video card is probably bad. Let us know what you learn.

Click to expand...

While this is considered minimum by a lot of people, i peronally think you should run prime for as long as your ever likely to run the pc at load also gaming rarely maxes out modern cpu's.

Just clock down to the regular clock speeds recheck seating on the heatsinks and take it from there.

The zalman also does a very good job of concealing dust just double check as mine often looks clean until take out the card.

Ill check it for dust again when I get home. I did turn the clocks back down to rated speeds to check, and it didnt crash as hard, but it was still locking up.

Oc1Kenube said:

While this is considered minimum by a lot of people, i peronally think you should run prime for as long as your ever likely to run the pc at load also gaming rarely maxes out modern cpu's.

Click to expand...

The reason I suggest this is that I've experienced several instances in which a failing RAM module could only be identified through several hours of Prime/Orthos testing. The PC would spontaneously BSOD when the faulty module was installed, even while idling, yet extensive Prime/Orthos testing was the only diagnostic that could pinpoint the problem. (Memtest86+ would not report any errors on either module) .