Another 'Too hot, or not?' thread for you

Sarcoth

Distinguished
Aug 14, 2007
2
0
18,510
Hoy! I would love some input on yet another temp-thread. I built me a brand new system during last weekend, boasting Q6600 and Noctua NH-U12F to cool it off. However, I'm a bit worried about my core temperatures and given the vast array of numbers being thrown around by random people on random forums, using random setups and random methods to calculate the 'true temp', I'd like some opinions about the numbers I'm pulling off.

As for case airflow, I'm using the stock three TriCool fans provided by Antec (came with the case, P182). Running the exhaust fans at low setting, while the one intake fan is set on medium. I'm trying to keep this as quiet setup as possible. Blasting away at high settings definately has an impact (of 4-6 degrees) on reported temperatures, but sounds like a jet-engine too.

I've settled to Abit's uGuru utility, CoreTemp 0.95 and EVEREST to check my temps, while stressing the cores with either Prime95 or EVEREST stability test. They all seem to be giving around the same values, so I'm pretty much doing all my diagnostics in EVEREST for now. SpeedFan gives really weird results, for example my CPU (tCase?) sensor is reporting negative one hundred degrees or so. Core temperatures are off that notorious -15C.

I have a slight overclock going, at 2880Mhz (320x9), stock vcore (CoreTemp reporting 1.3250V under full load). Running stock clocks drops the reported values by 2-3 degrees, CPU hovering around 55 under full load.

Idling around in Windows (3-8% load) gives me this, in EVEREST and CoreTemp;
CPU: 37
Core 0: 47
Core 1: 44
Core 2: 43
Core 3: 43

After about half an hour in EVEREST stability grind, this is what I get;
CPU: 57
Core 0: 74
Core 1: 70
Core 2: 68
Core 3: 68

Is this running too hot, or still very much in the safezone? System seems stable enough, although I haven't done any marathon stress testing yet. I used Zalman STG1 thermal grease with the HS, gave a good and thorough sweep for both, the CPU and heatsink, as instructed on the Zalman packaging. Apparently the stuff needs a while to settle in, so I'm hoping to see slight improvement in the temperatures in the coming days. I'm just worried the heatsink might not be installed as good as it should, maybe. The whole uncertainty and lack of trustworthy references makes me feel a bit uneasy. I really don't know what I should expect with this setup, and if there is more to be expected, should I blame the heatsink or airflow through the case.

Any input is highly appreciated. Thanks! :)
 

GijsJUH

Distinguished
Feb 16, 2007
28
0
18,530
Those temps are to high for that case, with that cooler, with that processor, and that overclock! For sure!
My friend has an Q6600 with an Artic Freezer Pro 7 installed, and runs it on 3 GHz, with temps not going higher than 67 degrees on the hottest core! And the Freezer isn't even running at max rpm! So maybe you could reseat the heatsink, and just apply a little layer of cool paste (I've also used the Zalman paste, and I put a thin layer on the processor *so I just couldn't see it anymore*, and the same for the heatsink... Than it's apply'ed(?) right...

And, since it's not much of an overclock, try to set the vCore to 1.3 Volts in the BIOS... This will sertainly help getting the temps down! (I think it's possible to go much lower than 1.3 Volts, but than you're sure of the fact that it will boot!)
 

mjb1206

Distinguished
Aug 6, 2007
10
0
18,510
No sure if it's fair for me to compare my temps, but just for the heck of it. I have my Q6600 B3 overclocked to 3.2, and am cooling with water. My idle temps are 44c, 43c, 43c, 43c, and my temps under 100% load are 58c, 58c, 60c, 60c..............
 

noisyninja

Distinguished
Aug 12, 2007
11
0
18,510
I wouldn't worry about it, plus you won't find any reliable answers here as it's just a virtual crapfest of people asking questions and the knowledgeable people not answering.
 

Mugz

Splendid
Oct 27, 2006
7,127
0
25,790
If a C2Q is two C2Ds side-by-side in a single package a la the Pentium D, then I'd imagine the information in the C2Duo temperature guide is as applicable. Nothing over 60ºC is advisable, nothing over 70ºC is safe, or something like that.

Leave your speed as is, just bring the voltage down as far as you can while remaining stable, and carefully investigate your cooling. Maybe upgrade it. Case cooling as well as CPU...
 

Sarcoth

Distinguished
Aug 14, 2007
2
0
18,510
Well, I went ahead and reseated my heatsink .. twice. First time around, it appears I applied TOO thin layer of Zalman thermal grease on the heatsink, and the base of the heatsink and my Q6600 'heatspreader' did not connect very well. Given the top-notch engineering quality on Noctua coolers (or that's what I've heard anyways), I suppose the Q6600 top is a bit concaved. That's not cool. Literally. Tcase temp skyrocketed to 58 in just under a minute of stress testing.

Next time, I spend over half an hour to perfect both surfaces with Zalman's STG1 grease, and applied some extra goop to the middle of the heatsink base and the CPU heatspreader. It appears it paid off, and I'm now at 51 degrees (Tcase) while stressing the CPU out with EVEREST stability test. That's four degrees less than before. Hopefully I don't go over 55 with a slight overclock. Then I'm just gonna leave everything as it is, and be happy about the results.

My abit IP35 Pro board does not seem to offer the ability to drop the vcore from 1.3250V, that's the lowest option in for vcore setting. Am I missing something?

I think 51C is pretty good, right? Please note, that I'm running both of my Antec TriCool exhaust fans at 'low' setting. This setup barely makes any noise, which was quite important factor when building my rig up.

Still getting pretty wild variation in core temperatures .. Core 0 running at 68 (100% load) while fourth core is stable at 63C.
 

TRENDING THREADS