Originally posted by: OILFIELDTRASH
I believe Im right. Hopefully someone can chime in and let us know.
That is a lot of ambiguity with Intel's thermal specification. But for the most part, those who have spent a great deal of time investigating the issue (myself and many, many others) have it pretty much boxed the issue as a non-issue.
First, the 67.9c 'Thermal Specification' does not mean the maximum temperatures of the chip can safely operate. Second, the 67.9c temperature it is referring to is not the actual cores, but the Intel Heat Spreader. This may have changed with i7, but back in the C2D days the only wany to know for certain the T-Case temperature was to get an engineering sample from Intel (they provide these to large OEMs) and make a hair line cut to the center of the chip on the top of the IHS and place a thermal probe. This is so that OEM's can design a product that will properly cool their CPUs. Third, the IHS (T-Case) tends to be (rumor, BTW) 10c or so cooler than the Tjunction temperature. So a the ball park figure was to take 10c off your average Tjunction temperature and sort of use that as your baseline of 'guestimate' for the T-Case. Fourth, the 'Thermal Specification' is Intel's 'average' result of temperature when climbing up to a power consumption of 130 TDP. Averages in general are just that. Not all the Core i7s hit 130 TDP under load and overclocked ones could great exceed that TDP.
Thermal Specification: The thermal specification shown is the maximum case temperature at the maximum Thermal Design Power (TDP) value for that processor. It is measured at the geometric center on the topside of the processor integrated heat spreader. For processors without integrated heat spreaders such as mobile processors, the thermal specification is referred to as the junction temperature (Tj). The maximum junction temperature is defined by an activation of the processor Intel® Thermal Monitor. The Intel Thermal Monitor?s automatic mode is used to indicate that the maximum TJ has been reached.
The i7 will have 5 diodes for temperature. It will have one in each of the cores and also another one on the geometric center on the bottomside, not to be confused with the topside measure that requires a special diode and hair line cut into the IHS.
I am not one to ignore my own findings and as such, I have ran 3 Q6600's beyond their 'forum' maximum temperatures now right around two years now. While the life of the CPU may be shortened with increased voltage and temperature, I think it will be a rare occurance for anyone to have a chip die on them before its usefullness has passed. Again this is not taking into account chip abuse. But for reference, I ran 1.4v on air on my Q6600 at 3.33Ghz. The cores hit 88c under prime for the B3, and the G0 hit 78c. For what its worth, information was published maybe a year ago that the B3 stepping actually didn't run hotter (much) than G0, but the Tjunction was not set properly for them. All these people (myself included, but I did question it) were thinking their B3's were running hotter than they really were. Truth is, they were only slightly hotter.
So, I mean, take some of this stuff with a grain of salt. Ask yourself this "Can I replace a $199 chip if it dies on two years?" Keep in mind, that chip may in fact be pennies on the dollar to replace later on if it dies. So, you know, calculate the risks and decide what you want to do.
BTW, more food for though, because I want to drive home that the issue was not black and white. The temp you see from your your trusted BIOS or CoreTemp may be significantly off. Check out more information
here at Tom's Hardware (keeps jokes out of this!)
Facebook
Twitter