High temperature of GPU1, Missing or inadequate ?

Having problems with Far Cry 2, crashes randomly. After much research I found a topic saying that these are random crashes of the Ethernet ports and that was enough to disable them. In fact, disabled my network card but still cannot.

So I thought that my 4870X2 is known to heat a lot which may be reason, despite the cooling of origin holds two slots on its own plus one for the card is 3 slots occupied just for graphics card!

Knowing all this, I met occt and I started a tested 5 minutes at the same resolution as the game (1920 x 1200), leaving it to Everest to monitor my temperatures.

I saw that GPU1 temperature rises to cruise in 3 minutes and then stabilize at 110°! It shows also that it is the only one to be as high and the other temperatures are relatively clustered. This brings me to the question: is it a failure of the card? or just the cooler is not big enough?

It is the VRM GPU1 tone that rises in temperature. Probably a defect in design or assembly. The VRM is near the power of the graphics card. It is true that its overheating tends to create artifacts, and may even lead to irreversible damage to the card.

In a first step back service because you already had artifacts, so the card has problem. Then replace the cooling system of the four ...

All the cards bi-GPU, I do not surprise me even more than that as Heater not allowed (and that consumption + Watts is EDF, which is content). For a smaller gain in games ...

Thank you for your answer!

I agree with your reasoning, except for damage. I have artifacts in Far Cry 2 that whereas the burn OCCT of 3, FEAR 2 or Crysis Warhead not.

In fact, I will try to mount the waterblock before the return to service, I inquired whether the disassembly of the cooling system was irreversible as the Palit 9800 GTX on and it appears not. However, I still square pad Gray ATI, so I can test without losing the warranty. M'enfin ... Unless I make a blunder

Good luck then, hope you recover well