I am posting this help request for I have not found a solution yet, neither on Tom's Hardware nor on the rest of the internet.
The problem is that the cores of my Q6600 have temperature disparities of around 10°C. To be precise, cores one and two are about 10°C higher than three and four.
At first I thought that's simply because the first two cores suffer most of the stress, but even with Prime95 and all four cores at 100% workload the difference in temperature persists. My CPU cooler is a Noctua NH-U 12P and I am using Arctic Silver V heat sink paste.
I am sure that the paste is spread correctly and the cooler is in the correct position. I have read about the same problem in another forum. The guy simply loosend the screw on one side of the cooler in order to release pressure on the first two cores. I am not sure what to think about that. Should I try that? Or is this solution simply an indication for wrongly spread heat paste? Anyway, I would need to know on which side core one and two are located on the CPU. I suppose they are on the left, but I am not sure. Have a look at this picture which I took from the instructions of Arctiv Silver V:
So what do you suggest? Am I right about the fact that 10°C difference in temperature on 100% workload is not normal? Should I reapply the heat paste? Or try to loosen the screw on one side of the cooler? And anyway, where are which cores located (1+2 on the left, 3+4 on the right?)?
Sorry for asking again, but I want to be sure. I feel like having put this thing into pieces and together again for one hundered times already.
So you think 1,2 are on the left and 3,4 on the right, don't you?
Thanks a lot man!
I have the same problem on my amd 6400 BE and I'll tell you it has nothing to do with your paste , I believe that the 10c higher core is acually reading the cpu temp. not the core temp like it should , that's my theory , my cpu 's performance is not being held back with the incorrect sensor reading and if you don't rma your cpu then you'll have to live with this wrong core reading , it's a common problem with intel and amd and some think the wrong reading is because of the placement of the core sensors , I don't but some do , so don't go crazy with the paste and if you can't rma it then dont worry about it and in the future I no I won't keep a cpu with the wrong readings just because it's incorrect , I hope I save you the work of r+r ing your cpu a few times !
because your dual core is on 1 die where as this quad core is on 2 different die. there could be many factors that will affect them. like the pressure being greater on oneside due to gravity pull,platness of the heatsink base,flatness of the IHS and heatpipe orientation. its just not as simple as you think mate!
Well, I disassembled the cooler and put it together again applying fewer heat paste. It's a bit cooler now in general, but the difference is still there.
Right now in idle its 40 - 38 - 34 - 34. I think I can live with it, for even with 100% it does not cross the 60°C line, and it's running OC at 3,1GHZ.
I suppose I can live with that, although I would like to have the issue solved... Well, one cannot have everything.
Maybe I will recheck the heat paste in some days.
Thanks for your advise. :-)
February 7, 2012 5:59:47 PM
I have 1 core that consistently reads 10 degrees higher than the other 3. Only this one (after a few hours) also scores a max temperature around 150 degrees Celsius. I can see it is a bug in the way the temperature moves. Gradual temperature changes look different from random numbers. This core reads from 60 to 150 and back to 60 in less than 1 second. It doesn't look very realistic. The computer works fine otherwise.