Well i've decided to try to push my Q6600 to 3.6ghz 1.45v, i originally had it sitting at 3.445ghz. I did the original OC a while ago, but i didn't use Coretemp to read the temps, instead i used ntune. Yeah i know ntune is blech, but i didn't know that at the time and the OC was stable.
Now that i've got it clocked at 3.6ghz i decided to try out Coretemp. What surprised me is the difference in temps between my cores. Core 0-1 are running hotter at 57C Idle, and cores 2-3 are running at about 50C Idle. I didn't think too much of it, till i fired up Prime95. The first 2 cores hit 83C in the first minute, while the other 2 barely hit 75C at their peak.
Before i fired up Prime95 i thought the descrepency at idle was due to the windows processes taking utilizing the first 2 cores a bit causing them to be a little hotter. But u'd think with all cores being fully loaded that their temps would be relatively the same.
The opinion i'm looking for is, what could be causing this? Could the weight of my Tuniq Tower be pulling down so the first 2 cores aren't getting as much contact as the last 2 (assuming the first 2 cores are located towards the top of the case)? I'm hoping that someone may have a better idea on what is goin' on here.
Ambient temp is atleast 22C in the room.
--------------- 14591 3DMark06 6210/6104/4507 Vista 64bit
Coretemp reports slightly higher temps than actual temperatures.
However I think your still running too hot.
Read the temperature guide by Computronix at the top of the overclocking forum. That will give you a better idea of how to calibrate utilities so your not guessing what your temperatures are.
I don't think the Tuniq is leaning because of it's weight. You might have uneven pressure on the thumb screws however.
Also what stepping q6600 do you have? A b3 will run dang hot at that vcore and frequency.
Message edited by sportsfanboy on 04-23-2008 at 03:03:36 PM
Ok that a little comforting to know Coretemp reports things a bit high.
Yeah, i decided to revert to my 3.45ghz config for the time being. I also tightened up the thumbscrews on the Tuniq and it seems to have helped some.
Ok i'll check out the Temp guide, ty.
The stepping is a G0. Had it a been a B3 i would have expected those temps, perhaps hotter even.
Also for reference, Coretemp (ATM), is showing 50,50,43,44 at idle. That's at 3.45ghz with a 1.4188vCore. nTune shows 46C which looks to be an ave of the four core's temps.
Edit: Just did a Prime95 test at 3.45ghz 1.4188vCore as additional reference. It's actually goin' atm while writing this. So far Coretemp shows each core at 77,77,69,69 at it's highest i've seen. According to the article i should not pass 65C to stay in the safe range. I'm wondering how on earth i'm gonna achieve that.
Message edited by lucuis on 04-24-2008 at 02:22:55 AM
--------------- 14591 3DMark06 6210/6104/4507 Vista 64bit
Core Temp reports temperatures on the high side, but I don't think even in Core Temp you should be pushing 80c.
I'm not looking at the guide as I type this, however I believe you might be running too hot still. 77c would make me nervous.
The deal is... The true tjunction (max core temperature) on intel core 2 is unknown. Intel's thermal specs are for max tcase, which is the sensor between the two dual cores for the heat spreader on the cpu. I know lots of programs say what they think the tjunction is, and it's a guess really.
highly recommend the temperature guide by computronix. That type of process is the only real way to know your temperatures, as all programs that I know of report them incorrectly untill properly calibrated.
Its not as simple as coretemps reporting your core too high. THe reason it might appear like this is because core temps monitor the actual core, where most other monitoring programs monitor the sensor between all your processors, where there is barely any heat compared to the actual core. For example, imagine two boxes. [ ] [ ]. Core temp monitors the two at their cores, i.e. [ * ] [ * ], whereas the others monitor it here: [ ] * [ ], so you get better measurements for the CPU with coretemps. DOnt sell it off as inacurrate.
---------------
Exile- Kreia what are you Jedi or Sith?
Kreia- Does it matter? Of course it does; such titles allow us to rip the galaxy into two halves: light and dark, good and bad. Perhaps I am neither, and I hold both as what they really are: pieces to a puzzle.
It doesn't matter. Those temperatures are far too high no matter how it's reported. I get 57/57/55/55 on coretemp at 3.6ghz, 1.45vcore, full load prime95 with air cooling. 70/70/68/68 at 4ghz.
Q6600's thermal limit is 71, at which temperature it begins to throttle back.
Its not as simple as coretemps reporting your core too high. THe reason it might appear like this is because core temps monitor the actual core, where most other monitoring programs monitor the sensor between all your processors, where there is barely any heat compared to the actual core. For example, imagine two boxes. [ ] [ ]. Core temp monitors the two at their cores, i.e. [ * ] [ * ], whereas the others monitor it here: [ ] * [ ], so you get better measurements for the CPU with coretemps. DOnt sell it off as inacurrate.
Darth I know the difference between tcase and tjuntion... "Core temp" reports core temperatures to high thats a fact I learned from computronix.
I'm pretty sure he knows more about this stuff than you and me.
It doesn't matter. Those temperatures are far too high no matter how it's reported. I get 57/57/55/55 on coretemp at 3.6ghz, 1.45vcore, full load prime95 with air cooling. 70/70/68/68 at 4ghz.
Q6600's thermal limit is 71, at which temperature it begins to throttle back.
What you need to understand is... 71c is the max tcase value not tjuntion (core temperature) for your go q6000. It will not throttle at 71c tjuntion(it will be running hot but not throttling hot).
What you need to understand is... 71c is the max tcase value not tjuntion (core temperature) for your go q6000. It will not throttle at 71c tjuntion(it will be running hot but not throttling hot).
The difference is 5 or 6 degrees at most. He hit 83 degrees C within the first minute of prime95. That is not a good sign.
Can someone help me also please i didn't want to start a new topic for the same issue
i also got a Q6600 G0 and i have a XfX 680i SLI updated to the lastest BIOS and my temps are very high for some reason
When i'm on Idle im at like 49C and under full load at about 62 (if i leave it for a couple of hours under prime95), i use the stock fan. I even tried at least 10 times to remove the fan, clean up the fan and cpu with a nice little towel, replaced the thermal paste and its always about the same temperatures.
I ordered some Arctic Silver 5 to try to lower the temps. i've finally received it, nice, i've managed to lower the temps from 49 to about 45-46(IDLE) and 62 to 56-57(Full Load) but even that is hot in my opinion.
Also im not sure if this is linked, but i've been trying to overclock a little, i've just wanted to put my CPU from 2.4 to 2.8 so i've read all the guides and stuff that were saying disable this, its a power saving option, disable that, and i can barely boot it at 2.6 Ghz but its not that stable, and 2.8 it just doesn't boot.
like i said i disabled the power saving things, i've set all the voltage options on auto, i've set the memory timings on auto, i've done everything...
-Anyone has any tips to share please ? or maybe a recommendation for another Heatsink/Fan for my CPU?
Just to mention, for the aiflow of the case, i've got a Antec Nine Hundred case, and i'm in the basement at around 21C
At this point i'm considering remounting the Tuniq. I'm also using the thermal paste that comes with the Tuniq, is this good enough?
Also is there any "best" way to mount the HSF and applying the thermal paste?
I've read a lot of articles and seen some techniques etc. But more opinions/tips are welcome.
Another thing i'm gonna do is improve airflow in the case. ATM some cords, i think, are restricting things. I'll take a pic if u all like, so you can all add some input.
update:i mentioned earlier my ambient temp in the room was atleast 22C. This was an estimate, i brought down a clock with a temp feature and it says room is 26C. I'm gonna guess +-2C since it's a cheapo thing.
Coretemp shows me at 52,52,45,47 with the 26C ambient.
At this point i'm considering remounting the Tuniq. I'm also using the thermal paste that comes with the Tuniq, is this good enough?
Also is there any "best" way to mount the HSF and applying the thermal paste?
I've read a lot of articles and seen some techniques etc. But more opinions/tips are welcome.
Another thing i'm gonna do is improve airflow in the case. ATM some cords, i think, are restricting things. I'll take a pic if u all like, so you can all add some input.
update:i mentioned earlier my ambient temp in the room was atleast 22C. This was an estimate, i brought down a clock with a temp feature and it says room is 24C. I'm gonna guess +-2C since it's a cheapo thing.
Remounting won't hurt. Give it a try. The thermal paste matches the AS5, according to Anandtech.
For me, I like to put a line of paste down on my Q6600, right in the middle. Then I let the heatsink do the spreading. Some people will object to it, but I like it this way since I found out that spreading it with something doesn't fill up all the holes.
--------------- "Nvidia, the Way It's Meant to be PAID Played! - Corrado
*Lesbian Lover Club* - founder Assman
One thing i didn't do that i see people doin is cleaning off the old stuff with rubbing alcohol. I just wiped the old stuff off with a piece of paper towel. The stuff i rubbed off was from the stock HSF.
Should i expect an improvement if i properly clean it off this time around? And how would i go about doin it? Could it be a simple as dampening some paper towel with rubbing alcohol, and gently scrubbing it off? Or is there something else i am missing?
Wow, just realized that was an onslaught of questions. TY in advance
One thing i didn't do that i see people doin is cleaning off the old stuff with rubbing alcohol. I just wiped the old stuff off with a piece of paper towel. The stuff i rubbed off was from the old HSF.
That's not good. Of course you're suppose to completely remove the old paste by using something like rubbing alcohol, using a lintfree cloth. This way you make sure there's nothing that can potentially decrease the surface area.
--------------- "Nvidia, the Way It's Meant to be PAID Played! - Corrado
*Lesbian Lover Club* - founder Assman
Also i mentioned earlier that i'd post a pic of the inside of the case.
As you can see i got a nice ratsnest of cords there i'm gonna try clean up. There's a fan in the front of the case which is the intake, it's pretty much blocked by the mess of cords, and the Hardrive is nestled in there too. I could try and move the Hardrive down to the very bottom bay on the edge of the screen there. (provided the cords will reach)
The bottom disk drive is a dud which i haven't removed yet. I'm thinking i'll remove that and put the working on a section down. Leaving room for a fan to be "rigged" above it to intake are. ATM i think i lack intake of cool, sort of, air.
Any thoughts or suggestion are greatly appreciated. I want to get this all planned out...like i thought i did orginally, before i start
14591 3DMark06 6210/6104/4507 Vista 64bit
i didn't want to start a new topic for the same issue 
and it says room is 26C. I'm gonna guess +-2C since it's a cheapo thing.
