A program like mem86, but for GPU's VRAM?

Here's the deal, I recently have two cases of computer "crashing"/"freezing" and then reverting back to "normal", except with the graphics in the game I was running all messed up (Warcraft 3 looked really messed up with totally screwed up textures on everything. I don't know if this is a GPU issue for VRAM issue...).

At the time, my GPU was downclocked by 25Mhz (Yes, downclocked), but my memory was overclocked by 75Mhz. My laptop specs are:
Dell Vostra 1500
2GB RAM
nVidia 8600M GT 256MB GDDR2 (Default core clock of 475Mhz and 400Mhz for memory)
Windows XP SP2

Before I updated my video driver (I'm on 177.41) from 169.28, I'd occasionally have nvcpl.dll crashing when I quit from Warcraft 3, stating invalid memory address (which could potentially mean corrupt video memory). The error went away after I updated my driver, so I didn't give it any second thought until yesterday when the crash happened, and today when it happened again.

So, does anyone know any program that scans the VRAM for corrupt memory? I still have 3 days (literally, runs out on the 26th of this month) on my warranty (It's a drop-in depot warranty, so I'll have to send it in...), so if nobody has any suggestions, I might just as well call them in and send in my Laptop to have them replace my GPU (but I honestly don't want to do that since my fall semester is starting, and I'll need my laptop since I'm taking Comp-Sci/Programming related courses).

To recap:
Core was DC'd to 450Mhz and Mem was OC'd 475Mhz at the time.
Warcraft 3 "freezed" for a second, then when it was unfrozen, all the textures were messed up (Seems like a memory related problem).
Therefore, I need a program to scan the GPU's video RAM for corrupt sectors. Something like the famous mem86, but for GPU.

Any help and/or suggestion is much appreciated!

 

No, there is no test programs for VRAM. Since you ask, I assume you have run the regular memtest86+ for system memory, which might also be corrupt?
Do you get these issues in other applications or games/3DMark etc.? Maby the card is overheating? Maby its memory chips are running too hot? Maby they don't take good contact to heatsink? You might want to open the laptop and tighten the screws. And while you are at it, clean the insides with compressed air.
To get atleast some idea through software, leave HWMonitor running in the background to see the max CPU and GPU temperatures after gaming. Also, Here are some examples what kind of images defective graphics cards can produce.
And have you updated all drivers, not just graphics card, and Windows+DirectX ? You might also want to run CHKDSK /r, to check potential disk errors, that could cause driver corruption.

 

Ronkotus, there are test programs for VRAM..

Just google "video memory stress test" and you´ll see...

 

Hey, thanks Samot, that's actually new to me. That must be the only test program for VRAM and released some time this year. I used to search for similar software when my video card was producing artifacts, so I didn't bother googling it now.

 

For checking for artifacts ATi Tools is a good program

 

Yes, try ATI Tool:
http://www.techpowerup.com/atitool/
It will stress a GPU extremely hard - click "Scan for Artifacts" after you have it installed.

 

Yeah, that was the one I used when my card was first overheating in a game and still use it to check that cooling is working. But that Video memory stress test might be better for checking faulty VRAM, not just overheating VRAM. Or maybe they just do the same thing?

 

Okay, I guess I was wrong--but my GPU is defective nonetheless.

The artifact that I saw in WC3 was like this one:
http://www.playtool.com/pages/artifacts/badgpu.jpg

Which, according to the article, is due to a bad GPU, not VRAM!

And since my GPU was actually downclocked by 25Mhz when this happened, it's definitely a defective GPU! The temperature was around 61C when I was playing, the highest it got was a few days ago when it was hella hot, it was around 75C... Does anyone think maybe the card sustained damage when it was really hot, and it started showing the damage just now?

Anyways, I'm going to call up Dell and see what they say...

Thanks for all the help so far guys .

 

75C is not very hot, more along the lines of warm. But yes, you definately have a bad GPU.

 

If you haven't delete the original Dell Utilities partition, one of the many diagnostic tests is for the video card memory which the test will scans for defects of vram.

 

Whoa, really?!

Either way, I just called in Dell, and they're sending me a box... Plus, I already confirmed that it's defective via ATiTools. It literally crashed my comp when I set my clock to 475Mhz/475Mhz (Core/Mem) and then pressed on "Stress Test".

Wasn't too cool when it happened, but at least I know for sure it's defective! Haha... I think I damaged it when I initially clocked it too high (520Mhz mem clock XD)!

 

Don't tell that to DELL, because if you do, then kiss your warranty goodbye

 

Heh, I'm smarter than that .

I didn't tell the Tech Support any more than "My GPU is defective, please send me a box so I can ship the unit to you guys to fix... Pretty pretty please with many cherries on top of it?"

 

What a nice guy you are... You also propably know, that you should remove all traces of these stress/OC programs from your system.