With that said I wouldnt bother with buying a temp gauge. If the HS is cool to the touch under load with a fan, then its cooling it. And if its getting hot with no fan, it is clearly pulling the heat up. I think your sensors got fried and are mis-reporting. And the fact its stable, rock stable, is proof that there are no problems. Save your money on the temp gauge, reseat the HS again and let it be.
As someone on the AMD forums used to say a few years ago, "If it's stable, then it's not too hot!" If you start to get instability, then a new CPU would seem to be the best way to go to me. Considering the amount of time and money that you are spending on this problem, then I'd probably go for the new CPU anyway... Andy
That is pretty much the conclusion I am coming to. It simply must be mis-reporting . . . there is just no way it can be producing the heat it says it is producing and still be completley stable. What bothers me is all the diodes must be doing it, because all 4 cores are way too high. However, based on previous experience from an ancient blast furnace of a P4 i used to own, once you pass that optimum thermal envelope wierd stuff starts to happen. By all accounts I have already long since passed that envelope and this machine is still all good. I have received the new TIM and will re-mount when I have more time tomorrow. If that fails to alter the temps, I am going to set the mobo to ignore the temps and overclock the @*%$ out of this q6600 to punish it for its sins . . . if it is indeed running as hot as it says I will know fairly shortly when I smell burning . .
Why don't you just return it as being faulty and get one that works, that way you know when youve OCed it too high. If its still under warranty that is.
Sadly the warranty on this puppy has looong since expired . . . besides the fact that I have abused it relentlessly in my quest for Ghz . . . Re-mounting the heatsink is going to have to wait until tomorrow . . amazing how time vanishes. .
Well as long as it does not crash, these chips have a killswitch of sorts if they get too hot, so not to worry. I hit 93oC on my Athlon 2700+ XP stock cooler stock clocks...! You have an ace cooler so everything is probably ok, the chip is a little fried. Try undervolting in the BIOS just incase
LOL This is why im afraid to do some ocing or dicking about with passiveness. I currently idle at 55 c
Well . . . I have finally found the time to remount the heatsink for a third time when my plans to go on a drunken rampage after work fell through On the plus side . . . my q6600 is now idling at a cool 32°c . . . The problem was with the mounting of the heatsink, but not in a very obvious way. The heatsinks on my IFX-14 are bowed slightly down below the top of the heatsink plate. On the right hand side, one of the heatpipes possibly due to a manufacturing flaw, bows a little lower than the others. What this means is that when I tighten the screws connecting the heatsink to the metal legs which attach the whole assembly to the motherboard backplate, one heatpipe is touching the metal leg. As I tighten the screws this side feels like it is getting very tight, but in fact all it was really doing was tightening the heatsink onto the leg as opposed to pushing it down onto the processor. Effectively what this meant was that the right hand side of the heatsink, while still touching the processor, was not tight enough. This allowed some of the arctic silver to very slowly ooze down the processor face when my case was vertical. When I took the heatsink off the processor the right/centre right of the cpu had lost most of its arctic silver due to ooze creep. This had not affected me on the old motherboard because I had mounted the heatsink in such a way that the legs could not touch the restraining screws plate. You can of course imagine my horror when I realized that my cpu had not actually been misreporting temps . . . but had in fact been running rock solid at 80 - 90 degrees I could kick myself for letting it go on this long, but I was pretty sure I had mounted it perfectly the 2nd time . . . Oh well, live and learn, thanks to all who suggested solutions . . . now I am going to overclock the s*** out of this ****
