Q6600 Overclock Advice

[DaSinical1]

I decided to upgrade my e6600 to the q6600 since microcenter basically forced me to =)... at least thats what I told my wife. Anyways, I bought the processor yesterday & threw it in my system. Currently I have it running at 3.6Ghz (400 X 9).

After 1 hour of running prime & atitool together... Speedfan says my temps are: SYS 29c, CPU 59c, C0 40c, C1 40c, C2 40c, C3 40c but Coretemp says my temps are: C0 58c, C1 55c, C2 54c, C3 56C. I'm gonna agree with Coretemp & the higher temps just to be on the safe side & since its running under 60c I'm assuming that I am good to go temp wise. I am currently running a lapped TRUE as my heatsink & Shin Etsu TIM. At idle Speedfan says my temps are SYS 28c, CPU 40c, C0 22c, C1 22c, C2 22c, C3 22c & Coretemp says that my temps are C0 37c, C1 36c, C2 32c, C3 37c.

Also, both speedfan & cpuid show that vcore to be 1.36 under load. Vcore drops to 1.34 at idle.

Do these look like good enough setting to run 24/7? I'm gonna let the atitool & prime run at least 8+ hours to ensure that the oc is stable but so far so good.

Mark

 

[krnmastersgt]

Sounds good to me, the rule of thumb when using speedfan on a quad core is to add 15c to the temps it displays for the actual real temp. I prefer to just use CoreTemp and call it a day however, fine temps for running 24/7.

 

[Tempered81]

Excellent 58C on load at 3600mhz on a quad. I'd say that's great. Im sure the shin etsu under a lapped TRUE has something to do with that

 

[krnmastersgt]

Someone should do a comparison, get some Shin Etsu, AS5, ICD7 and use the exact same temp controlled room + chip + cpu cooler and see which gets the best performance, I'm hoping the ICD7 I just got outperforms my AS5.

 

[DaSinical1]

Krnmastersgt, i have both Shin Etsu & AS5... I don't have the icd7 though. On my previous e2180, e2200 & e6600 the Shin Etsu consistently got a few degrees cooler then the AS5 on load temps. I tested the AS5 by spreading evenly like i do with the Shin Etsu as well as just placing a drop (half the size as a grain of salt as well as full size a grain of salt) & the Shin Etsu consistently got 2-3c cooler. As always YMMV but needless to say that I will only use the Shin Etsu on my computers but have no problem using the AS5 on computer that I build for others. I am very curious about the ICD7 now but after putting the Shin Etsu on the TRUE, northbridge, southbridge & accelero... I don't think I wanna take the time to experiment anymore unfortunately. Maybe in a few months when its time to cleanup the case & re-apply the TIM.

 

[aigomorla]

Originally posted by: krnmastersgt
Someone should do a comparison, get some Shin Etsu, AS5, ICD7 and use the exact same temp controlled room + chip + cpu cooler and see which gets the best performance, I'm hoping the ICD7 I just got outperforms my AS5.

lemme just add something.

I was on my ever so quest for the perfect TIM. Ignoring everything which my friend yoda was telling me on how all tims are simular, i still went off on my arogant way to find it.

I can safely say i have tired probably almost every TIM out there. From the toothpaste looking crap that comes in zalman's heat sinks, to the mystery white stuff that comes on evercool chipset coolers, to the ever so popular liquid metal by cool it, <trashed tht right away as it melted alu foil>, to IC DIAMOND, by FAR the hardest POS to apply on.

My alltime favorate is this:
http://forums.pcper.com/showthread.php?t=415964

I recomend GBR as a source as well if you guys cant find it cheaper. I have bought TONS of shinetsu from him, and he has always delivered. Even lost a vile, and he offered to replace it for free.

Shinetsu by far is he easiest to apply. Just blob and top. Both IHS and Die. The stuff is non conductive, so its okey if it tops off a little. And i have the best and most painless application with it.

 

[krnmastersgt]

Well the ICD7 might have been hard as hell to apply, but how different were the temps between it, AS5, and Shin Etsu? o.o

 

[VirtualLarry]

I hope I have such good results with my MC Q6600s. I'm going to buy some Tuniq towers to go with them. I don't want to have to lap anything.

 

[myocardia]

This is pretty good reading, if TIM's interest you. It uses the most poular, highest performing TIM's available in the US. Unfortunately, it doesn't use Shin-Etsu, although most people seem to think that Arctic Cooling's MX-2 is just relabeled Shin-Etsu.

 

[Shooks]

So @ 3.6Ghz is 58-62 reasonable under full load? Those are the temps I am getting with a Nirvana 120

edit: Wow, this chip just keeps impressing me. Dropped the voltage to 1.46 and now it maxes out at 57
And I just read the OP again and noticed even lower voltages, seems I need to try some lower values.

 

[kkk60091]

ICD7 stained my lapped Ultra-120 ex and unlapped qx9650. left some kind of watermark on the both surface. I cleaned them with arcticlean 10+ times. still a mark left. I end up relap my heatsink and lap my qx9650.
No idea why it was like that. I trashed all three tubes I bought from frozencpu.


currently using TX-2, used Shin Etsu X-23-7783D for few weeks. the performance is kind simmilar to me under prime. ( well, I didn't collect data to compare, just impression).

btw, i don't think they are the same stuff.

 

[cpemma]

Originally posted by: krnmastersgt
the rule of thumb when using speedfan on a quad core is to add 15c to the temps it displays for the actual real temp.

That's for any CPU with Tj=100C, Speedfan uses Tj=85C in the calculation so 15C low. Try the 4.34 beta, the bug claims to be fixed.