How Do I Know if My Video Card is Bad?

Modern graphics cards possess GPUs which are among the most powerful processing units ever to make their way into a desktop computer. For gamers that is mostly good news, as it means that finding a capable graphics card for a low price is quite easy.

But there is a dark side. The cutting-edge performance demanded from video cards means that Nvidia and ATI up the speed of their products as much as possible without causing obvious instability. That can cause problems down the road. How do you know if your graphics card is about to kick the bucket, and how can extend the life of your video card?

When it comes to prolonging the life span of any piece of electronics, it is best to be proactive. The early signs are not malfunctions, but simply represent an environment which is hostile to the life of a video card.

The first sign of future video card problem is a high GPU temperature. There are several different programs that can check the temperature of a GPU, but I prefer Rivatuner. There are some video cards available today which will in fact run at temperatures as high as ninety degrees Celsius and not see immediate corruption or failure. That said, ninety degrees Celsius is extremely hot and it certainly can't be good for the overall lifespan of the video card. Keeping things below seventy degrees Celsius is probably best. This is not a scientific number, as there has been no through testing ever done of how long video cards last under different environmental conditions. It is a number which most video cards can sustain, however, while it is in the ballpark with maximum temperatures one would want to see on a CPU or a motherboard's northbridge.

Another related early sign of future problems is a fan which seems unusually loud, rattles, or seems to waver in speed. A fan which is not working well could result in higher than desired temperatures. It could also fail, which can certainly cause problems.

General instability is perhaps the most serious sign of future problems. A video card which seems to regularly crash when it comes under load means that the card is moving beyond the limits of what it can endure, resulting in errors which cause the graphics driver to become non-functional, resulting in either the crash of an application or, in some rare cases, the entire operating system. While high temperatures and a defective fan hold the potential for problems, instability shows that there are already problems happening, although they may not yet be fatal.

If you're wondering if you're video card is bad because it is outputting artifacts, distorting images, or otherwise messing with the visuals that it should typically display, it may be too late. Video corruption is the number one indicator that there is something wrong with a video card, and by that point it may be too late. On the other hand, it can be difficult at time to decide what is and isn't being caused by video card corruption, particularly if one owns only a few games. It could be some other system malfunction causing the issue.

There, however, a way to test for corruption is probably to use a program called FurMark. FurMark is simply a stress-testing program like those used to test overclocked CPUs. It taxes the GPU as much as possible in an attempt to cause instability. It is important to note that many video cards will show a few artifacts when running this program. A video card on the fritz, however, will likely show an extremely large number of easily visible artifacts. It may also cause the program to crash or cause the video driver to malfunction. Note that there is some small amount of risk, as stressing a GPU which really is near death could push it over the edge.

There are other signs that a video card is going bad, but they're fairly obvious. A video card stalked by death may work only intermittently, may not be able to launch 3D applications at all, or simply will stop outputting video altogether.

So you've run the tests and you've found some signs that you video card is worse for wear. Is there anything that can be done?

That depends on how severe the problems are. If one is simply looking at the early signs of video card problems, the resolution is fairly simple - run the fan more quickly, which can be accomplished with Rivatuner. Alternatively, it could be helpful to improve airflow throughout the case by adding more case fans or running them at higher speeds. This should help improve stability and will hopefully prevent failure.

If one is looking at the fatal sign - significant video corruption - then it is probably time to start looking at a new video card or contacting the card maker's warranty department. Although lowering GPU temperature helps decrease the problem, a card which is causes a lot of artifacts is probably well on its way out.