I've been having trouble running a few games recently. What happens is that, randomly after playing the game for a while, my computer goes into this weird state similar to BSOD without the BSOD. It locks up, the fans start cycling on and off quickly and the only way to leave this state is by holding down the power button. Off the top of my head, these games include Half Life 2, Portal 2, Cities XL 2012. Within the same time period, I have played LoL, WoW, SC2 and other games regularly without this problem.

Things I have done to try to solve this problem is that I have updated my video card drivers, my sound card codec (?), I've been running a temp monitoring utility as well. The video card runs at 65-66 degrees C when outside games and very often gets to 99 degrees C when in games. Again, getting to 99 degrees happens with a bunch of other games without any problems (that I know of).

I tried running the FurMark VGA Stress test and the computer locks up every time after 10 seconds. This basically proves to me that it is the video card. It seems to lock up the computer whenever the video card temp reaches 102 degrees.

The computer is relatively new (4-5 months old) so I'm not sure what is causing this. Here are some specs that might be a bit helpful:

Total amount of system memory 8.00 GB RAM
System type 64-bit operating system
Number of processor cores 2
Display adapter type NVIDIA GeForce GTX 460
Total available graphics memory 4095 MB
Dedicated graphics memory 1024 MB
Dedicated system memory 0 MB
Shared system memory 3071 MB
Display adapter driver version 8.17.12.8562
Primary monitor resolution 1920x1200
Secondary monitor resolution 1280x1024
DirectX version DirectX 10
Network Adapter Realtek RTL8168D/8111D Family PCI-E Gigabit Ethernet NIC (NDIS 6.20)

What do you think?
Hey Practo,

It definitely seems like the video card is faulty. It is failing during high loads which are causing high temperatures. Not all games, although they raise the temperature to the same level, utilize the GPU the same way. Therefore, some games cause it to lock up while others don't.

LoL, WoW, SC2 are less GPU demanding than Half Life 2, Portal 2, Cities XL 2012.

Since the PC is 4-5 months old, I'd go through the RMA process for the video card and have it replaced. Which brand is it?
Try resetting your CMOS and repeating your card.
Also, what is your PSU?
Try diagnosing system ram with MemTest86+.
Also, check your CPU OC (If you have OC-d).
Goodluck,if this doesn't solve your problem or if you have any questions don't hesitate to PM me.
Xsever wroteHey Practo,

It definitely seems like the video card is faulty. It is failing during high loads which are causing high temperatures. Not all games, although they raise the temperature to the same level, utilize the GPU the same way. Therefore, some games cause it to lock up while others don't.

LoL, WoW, SC2 are less GPU demanding than Half Life 2, Portal 2, Cities XL 2012.
Yes, although these are older games, they pull much higher framerates, thus causing a higher load. For example, starcraft II caused GFX cards overheating on the MAIN menu, simply because they didn't include a framerate limiter when at the menu, causing the graphics card to render in the thousands of frames.

Not all games raise the temperature to the same level. Since a GFX card is multiple units clumped together (shaders, cache, memory controller, RAM chips, etc...), a game may utilize the shaders at close to 100%, but not the memory controller. There are also different types of calculations. Hence even if they load the GPU to 100%, it's only one part of the GPU that is saturated to 100%. Now furmark shows GPU load to be less than 100% (it varies between high 80s and high 90s), but it stresses almost all the components, for example the memory controller (I have seen 60%), although it's a core / shader focused stress tool.

@practo: 102C is incredibly close to the thermal threshold that the GTX 460 can withstand (according to Nvidia, and set by default as overheat protection), hence it shuts down. The card will not last at those temperatures. 65-66 degrees is TOO high on idle. Do you set fan speed manually, or automatically? Did you OC, overvolt, etc...? Download GPU-Z and check clocks, temperatures, voltages, and post them here.

@AvoK95: How can the PSU impact the graphics card's temperatures? The card will accept voltages within the ATX specifications (11.4V - 12.6V), beyond those boundaries the GFX card is not supposed to function. On the crappy PSU I had, as soon as voltage dropped below 11.4V, the PC would restart. Since there are no power - related errors apparent (they are heat-related), then even if the PSU had a problem, it's not the problem at hand.
I think you meant reseating your card. Practo, do that, and if you have a blower, clean the card from any dust, then reseat it. Do not tamper with it if you expect warranty, but you can remove the dust, of course.
yasamoka wrote@AvoK95: How can the PSU impact the graphics card's temperatures? The card will accept voltages within the ATX specifications (11.4V - 12.6V), beyond those boundaries the GFX card is not supposed to function. On the crappy PSU I had, as soon as voltage dropped below 11.4V, the PC would restart. Since there are no power - related errors apparent (they are heat-related), then even if the PSU had a problem, it's not the problem at hand.
I think you meant reseating your card. Practo, do that, and if you have a blower, clean the card from any dust, then reseat it. Do not tamper with it if you expect warranty, but you can remove the dust, of course.
If it was heat problem the PC would just simply shut down and give a siren when it's turned back on.
Also, when a PSU has less amperage and wattage then the recommended specs it will cause such problems, I have had many moments where I changed the graphics card and the same problem continued, when I changed the PSU with a 300W Enlight (it was a 400W Chinese) the problem got solved.
Try updating the BIOS of the graphic card.
though it could also be faulty ram memory, I had a similar problem years back.
Evga gtx 460 strikes yet again...
Thanks guys for all your comments, I'm gonna go through the things you suggested. Just wanted to mention though that I ran MemTest86+ and everything looks fine there. Thanks again.
Is it just me, or am I really doubting that a batch of EVGA GTX 460s arrived to Lebanon, all with extreme cooling problems? This seems to be the same issue as Shant's, but with Shant, the fan speed kicks up to 100% then the temps become 75-76C.

@AvoK95: It's not a heat problem?? 99-103 degrees is not enough to cause lockup? It's enough to cause card failure.
Guys, the thermal limit of GF100 is 104, not gf114. Dont confuse both, alot off ppl online think that the gf100's limit is the same as gf114, it is NOT. Remember although its lesser in cuda core/rops/..etc, it is smaller in size. Matter of fact the 460s thermal limit is 95.

OP: Try and speed up the fan, open the case and try.

@ Yasamoka: Its not the batch sent to us, it is what it is, noisy and hot. People tend to avoid evga EE cooler these days. Also shows the weakness of our market in cases and fans. If we had good cases and a wide collection of fans we wouldnt be complaining now would we?
Your only choice is to RMA it (Sent it to warranty)
Beej wrote@ Yasamoka: Its not the batch sent to us, it is what it is, noisy and hot. People tend to avoid evga EE cooler these days. Also shows the weakness of our market in cases and fans. If we had good cases and a wide collection of fans we wouldnt be complaining now would we?
But I've never even thought that a product could be released that overheats by default, at normal room temperatures, with ample airflow! It's unbelievable! I would at least expect a cooler to keep temperatures at least within range, not too superb, but just fine, this seems to make the cards exceed the thermal threshold! And when I've tried with Shant, no matter how much thermal paste you had in there, it barely made any difference. The previous application covered almost half the GPU only, but in a pattern (not even a circle), when we reapplied, and covered the whole base, temps even went up by a degree (margin of error I know).
15 days later
Just an update since I've gotten a couple of questions. I contacted the company and will be sending it in for an RMA. Will let you know if the new one I get works.