Does anyone know what to do when your game suddenly starts artifacting? I had this problem today. All of a sudden, my screen would act up and turn all sorts of crazy colors (fallout 3 screen: http://img602.imageshack.us/img602/6154/sydneybug2.jpg http://img193.imageshack.us/img193/6020/sydneybug.jpg ) I had been playing Fallout 3 and FNV on the new NVIDIA drivers that were released on oct. 25th without any problems. All of a sudden, I can't play any game on my computer without this happening and then crashing to desktop with a 'nv4 has failed' or some such message. GPU heat never goes above 70C, anyone have any idea how to fix this?

NVIDIA 9800GT
AMD Phenom II X4 3.2ghz
4GB RAM
650W corsair PSU

Don't think you're getting accurate temps if the gpu never goes over 70c man. Could be a sign of your gpu failing. Are you overclocking it? If so lower the clock speed a bit.

When I have seen such artifacts in other games, it has allways been because of either Graphics card / CPU / RAM overheating.

Try setting all settings to minimum, and see if it fixes it.

Do you have your GPU fan set to a constant speed? I find that usually helps me keep my card cool, I keep it at 80% at all times and I never run above 70c. (This is with a GTX 460, overclocked quite a bit).

Another suggestion, try to reseat the GPU, that is to unplug everything, pull the card out, dust the slot and the card very well, and put it back together. Sometimes this helps stuff like this, but aside from this all I can think of is your GPU is either overheating or its going bad.

It's not overclocked, and the case has three fans. It probably isn't accurate because everest showed a much lower temperature than the other program I was using.


Question, what does GPU mean? Video card or RAM? Doesn't power supply have a role in artifacting as well?

Graphics Processing Unit, so yes, video card.

It can if it can not supply sufficient power to the GPU, but I suspect the GPU is at fault here.

Try uninstalling the GPU drivers then reinstall them. Could be driver corruption.

If the GPU has not been cleaned of dust in 2+ months, I'd recommend you blow out the dust with an air compressor or canned air. Heating issues can lead to what you are seeing.

Just curious, if it was the video card, wouldn't I be experiencing other issues as well? Could it be the RAM that I'm having problems with?

And yeah, I checked last night, it's not dusty at all. This card is 2 years old, but I just switched it to this machine a month ago without any problems.

2 years old without maintence there is bound to be some dust, some hair, you do shed skin cells after all, if you used auto fan on the gpu then the tim/thermal interface material/thermal compound has turned into a hard chalky/putty substance that is doing little to no heat transfering.

err, is it normal for a video card to just become useless after 2 years like this then? I've never heard of anything like that.

does not mean its useless, it just means its in dire need of maintence=) far from useless, evga puts out mails now every so often reminding its customers its about time for cleaning up their computers again. reduces rmas when computers are taken care of.

What materials would I look at buying to fix the whole thermal compound thing for the video card?

A little update...

http://img109.imageshack.us/img109/6065/screenshot28v.jpg

After nearly 6-7 hours of stress testing with prime95, and seeing no dangerous temperatures listed on everest, and also uninstalling/reinstalling the latest drivers for NVIDIA, I decided to fire up Fallout 3 and see what happens. As before, the game runs fine for about 3-4 minutes, then all of a sudden, the picture goes out of whack, momentary freezing, pic goes out of whack more, then the computer is completely frozen and I have to boot from the desktop button.

It's worse now, though. After restarting my computer, the artifacting started happening on my desktop and while browsing google - it's not just games now. Another restart afterwards (which im posting after right now) shows none of these problems, so I guess the problem 'cooled down' or something (still dont want to say that since Im not even sure if its heat yet).

For the sake of it, I loaded my BIOS before this latest restart to look at the temperatures, it was nothing serious at all. How can system heat be causing artifacting if my computer temperature isn't running too hot?

Any further help would be appreciated, I'm completely stuck on what to do now.

prime95 stresses your cpu and ram not your gpu. most likely gonna need a new vidcard.

The problem is in relation to your videocards memory. Chips are either too hot, or failing.

I appreciate all the responses, im just confused about one thing.

If the video card is failing, wouldn't I be getting major FPS lag before freezing or artifacts? Why would it end up going from running fine one second to freezing/artifacting in the next after 3min of play like that?

The GPU temp is in the 40Cs, it can't be heat.

Can you get your hands on another card and test?

This is my guess. Defective memory modules will cause artifacting, freezing, BSoDs, and general non-specfic errors.

If this is the case, the GPU may act normally until a bad memory chip is used.

Unfortunately, (from personal experience) once the memory modules go bad on a GPU, the GPU is trash (unless under warranty).

Ok, sorry all to keep this thread going I know it's veered WAY off from FNV but the information has been extremely useful to me here and just have a couple more statements/questions.



No high heat temperature or gradual stuttering/slowdown required? Its just like a 'scar' or something on a certain part of the card, where once it reaches that point, it gives out?



Hmm, thats a good question. I got the card in September 2008. Can't remember if it's under warranty...

Ok, here are the results of a FurMark test:

After 3 mins: http://img824.imageshack.us/img824/4436/63612001.jpg
At 5 mins, 56 seconds: http://img99.imageshack.us/img99/8828/74721013.jpg

The first one, it froze up only for a moment. Then it continued just fine. At 5:56, it froze for a long time, then went back to normal, then artifacted again at 5:59. At this point I got too worried to let it continue running. As you can see, both Everest and FurMark log GPU temps and it was high in both of them, even Everest which is supposed to be more accurate. After I stopped running Furmark (as I'm writing this post), I still got some artifacting but it wasn't serious, just 'blips' of it and they've mostly gone away now.

Based on these results, would you say the Power Supply is not giving the card enough power, or are certain functions on the card just straight up dying?

Usually if its a power supply issue you would get BSoD or other fatal errors required that you reboot.
You have serious artifacting so I can confirm a GPU malfunction (most likely memory). According on the Everest reading its not the GPU core that is overheating, but most GPUs don't tell you the temperature of the GPU memory. However, it doesn't really matter the GPU memory is defective, overheating or both.

If you have a warranty then get the paper work started for the RMA, if not you'll need to buy another graphics card.

Another update: Now, as I am browsing the internet, I get random occurrences of artifact lines / freezing. I can move the mouse just fine, but nothing on the screen responds. Then, it unfreezes and proceeds as normal.



Nope. I don't even have the box I got my gfx card in 2 years ago. I guess it's time for a new card.

Any suggestions for the 100-150 buck range?

edit: Also, does the temperature level in there suggest my case doesn't have adequate cooling? What conclusions should I draw from the temperature listing on FurMark?

Update:

I refitted and dusted off the video card (although it didnt have much dust on it at all..not dirty at all) and the artifacting was less apparent during things like web browsing, but still exists and is still only minutes away from happening when I play a game. This time however, it artifacted, froze for about 15-20 seconds, and then went to a BSOD for ~1sec before the computer auto-rebooted.

So...PSU or video card?

By the way, I couldn't get the temp when the game artifacted, froze, then crashed and rebooted but when I restarted I immediately opened up the software I have and it said the temp was 57C.

More than likely video card. Easiest way to find out is to grab another card that has roughly the same specs(and draw). Plug it in, and see what happens. The other option is to drop the $50 on a quality power supply(enermax, thermaltake, etc), and see if it continues to happen.

New update:

Tried using the 430W PSU in the same system. Same artifacting problems, so it wasn't my 650W corsair PSU causing the artifacting.

I'm having a new problem now. 9 out of 10 times I turn on the computer, it artifacts from the boot screen and doesn't even make it to the desktop, all I see is a black and white splotch with colored squares sitting there, then I get a 1-2sec BSOD and it restarts, and goes on and on and on like that until I force-power down the computer via tower button.

Before I go get a new video card, can I be sure, based on my tests its the gfx card and not the motherboard/RAM?

[censored], it's artifacting again as I write this.

if the RAM was defective, you would get random crashes/blue screens no matter what you do.
defective motherboards are rare and very difficult to diagnose.

all the images you showed indicate a broken graphics adapter.
and if you think about it, it should be very obvious, that it's the graphics card.
furmark running => more artifacts


check out some prices on the web or ask your competent forum folk for help on that one. =)

good luck with your new card. ~=)