Q6600 core temps vary by 9C

[Golden Dragoon]

I have just gotten a Q6600, and am not overclocking it just yet, but running prime95 and core temp is showing me that the cores are running quite different temps from each other at load, I am getting the following:

Core 0: 55
Core 1: 49
Core 2: 50
Core 3: 46

Now my "old" C2D used to show a small diff of <4C between the cores at load, and most of the time almost the same temps, should I be worried about the temps I am getting? I have tightened the screws, and made sure that they are all about the same tightness so I doubt that anything is mounted wrong.

Edit: looking arround these seem to be quite high temps for a stock C2Q anyway, so it could be a problem with my thermal paste application, but I did it just like I have allways done and used AS5, so I would hope that isn't the problem but if anyone thinks that this might be it I will whip it apart again and reapply.

[Thorsson]

That's a lot. Are you using the stock cooler? It seems possible that either your HSF bottom or the chip itself are not flat.

[Golden Dragoon]

Nope those temps are on water (LR Storm G4 waterblock + 120mm rad)
you can kinda see why I am fairly worried, but I didn't do anything different with the application of the AS5 (though it did seem thicker than I remembered it being)

I wasn't too concerned about the higher temps to start with, I simply thought it had something to do with there being twice as many cores, but having looked arround it seems that quads have very close temps to the duos in most cases, and that something in my system mustn't be right.

[Webby]

Don't forget the storm was never designed for use with quad cores if I remember correctly it is an impingment block if the jets are not hitting over all the cores you will see variable temperatures (core to core) it may be time to get a new water block!

[Golden Dragoon]

True enough, but it did more than well enough on a dual.

Anyway, reseated with AS5, and noticed my mobo was clocking it to 2.6ghz at 1.35v (2.4, 1.2v stock), but my temps have now dropped to 50/48/46/49 Load which is a much better spread, and after a while the AS5 should settle and drop those by a couple of C as well.

[Webby]

Yeah, the problem is that the cores of a quad are not centred, or at least all four are not directly under the jets so rather than on a single core the entire core being covered by the jets you moved to a duel which due to a smaller manufacturing process probably 2/3 of the cores were covered with a quad though the is a gap in the middle between the two sets of two cores so now you are probably only hitting a quarter/third of each core so cooling becomes more inefficient and your temperatures spread as different sections of each core are being cooled more or less efficiently.

[looney]

the temps on the four cores on mine also vary, for example at the moment core 0 42c and core 3 is 36c, and im just browsing the nets at the mo, so pretty much 'idle', i thinks its quite common, but im suprised ur loads temps are that high on water cooling, altho nothing near to fry ur chip , my loads temps are max 58c on an akasa 965 (which isnt seated properly i think, but cant be arsed to seat it at the mo).

[looney]

by the way how do you apply ur paste? the AS5 isntcutions are just to do a thin line in the middle of ihs.

[Golden Dragoon]

The G4 has a large jet impingment area, prob 65-75% of the quads are being directly hit by the jets, and prob close to 100% of a dual.

You can just about see it in the pics on this review: http://www.procooling.com/index.php?...s&disp=90&pg=1

The hole in the jet impingment block itself on the right is from what I remember just about the od of 1/2" id silicon tubing which gives you some scale to the ji area then the top of the block is where it narrows to the standard thread.

[Golden Dragoon]

by the way how do you apply ur paste? the AS5 isntcutions are just to do a thin line in the middle of ihs.

I do what I always do, I put a thin line on, then spread it with a razor blade, and remove any excess.
I found that doing that instead of AS' instructions was better for my C2D.

[Golden Dragoon]

Update, After completely flushing my water system and reseating my blocks my full load temps are now 57/56/54/58, that is with 3D mark running to stress the GPU as well as Prime 95, a much closer temp spread, and as the processor is now running stable at 3,708mhz (1.35V, 412mhz fsb) I find those temps much more acceptable, at idle the temps are roughly 32-35C, not sure how much higher this processor can go, as things get unstable at 424mhz not matter the voltage, I think it is the motherboard or the memory to be honest as I am pretty sure there are another couple of hundred mhz left in this chip with the proper tweaking, and lowering the memory clock makes it stable for a couple of hours, but it eventualy fails orthos, whereas with the memory clocked higher it fails in mins.

I did notice that after a short time my PWM temps were about 85C+, which I was worried about, but I though about it and realised that a waterblock won't put any airflow over the pwm heatsink so I fitted a 60mm fan to my waterblock pointing towards the pwm area which has dropped the temp down to a much nicer 59C