• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Another Q6600 Core Temp (?)

liquidfluidity

liquidfluidity

2[H]4U
2FA
Joined
Dec 28, 2006
Messages
2,938
Finally got my system up and running and all is well. I have the Q6600 clocked to 3.0GHz at my stock voltage(1.2v) which is actually about 1.14v after v-droop. Needless to say , I have excellent temps. My question is how far apart should the cores temps differ from one another and should that gap close when running Prime95? I'm using Coretemp and at load , I'm getting 42-37-42-36C(1/2 hour in). The cores never seem to get much closer in temps and they are all at 100%. Should I maybe try re-seating my waterblock or just not worry about it. I'm very anal about stuff so this is kind of eating at me. The thumb screws are all adjusted to the same exact tension as I measured threads with a machinist ruler:rolleyes:U I suppose the springs tension could be a little different but that's me thinking way too much!

Another question - when the cores are numbered(0-1-2-3) at what orientation are they on the di(e)? I'm thinking I may be able to adjust my tension by knowing what core is where.
 
Thats normal to have 5-7 degree variance between cores. Nothing to worry about. ;)
 
That's cool! I have noticed that one core always seems to run cooler. Maybe it's just the last in line to get any work;)
 
Wow. There's actually about 3-5*C difference in the 2 programs. Is Real Temp the trusted program of choice now?
 
Your clamping pressure should be uniform. I use precision die springs from mcmaster.com so I know the spring rate (provided on the site as part of the spring specs) and by knowing the threads per inch of the threaded rods I can dial in (fairly accurately) the clamping force I want. Sorry I didnt keep the part number, it took about 2 hours of searching through all the zillions of springs to find the one I wanted. Start off specifingy rod diameter and overall lenght and ends ground flat will thin out the selection.

You can damage the cpu with too much [pressure, Intel specifies the min and max package (cpu package) loads at 17Lbf min and 70Lbf max. I am not sure if the socket has any limitiations, I found that pressures over 50 lb force did nothing for temps and there was little change in the 30 to 50 lb range. The stress this puts on the motherboard is not trivial. It would be possible to "jack" the nuts used under the board through the mounting holes if not careful. I use faily thick nylon washers to help spread out the force on the backside of the board.

computer007.jpg



Sounds like you are like me, (sorry bout that) get some really good springs and then remount and if the delta is still there, dont worry about it.
 
I'm using a D-Tek Fusion with a backplate so the only real variable would be the springs. I counted threads and measured in mm to make sure all was even(appeared). When I installed the block , I set it to what the specs said which is 3 full turns down(I think-can't remember right now). Both the block and the chip have been lapped for flatness.
Sure , now I'll be thinking about springs the rest of the night!
 
liquidfluidity said:
Wow. There's actually about 3-5*C difference in the 2 programs. Is Real Temp the trusted program of choice now?
Click to expand...


there should not be, make sure the sampling times are set the same and what ever is used as Tjmax agrees between the same programs. Then they should match.

If they still dont look that the option and read the delta to Tjmax, this must agree between the programs or one program or the other is programmed incorrectly.

You should be looking at delta to Tjmax anyway, its the only measurment that matters.

Here is how you can read it directly yourself.

http://www.hardforum.com/showpost.php?p=1031080147&postcount=5

The comment at bottom about TAT is most likely incorrect and is due to the progam using a different sampling time and thus the average temp over whatever time peroid varies. But all these programs have evolved and had changes applied. This is why the detla to Tjmax number which comes directly from the cpu register is the number that is important. See the data sheet from intel on your processor and the Thermal and Mechanical guildelines for more information on " Delta to TjMax" , (or the 100000 posts here) .
 
Yep , just got to looking and Real Temp Tjmax is set 5C cooler. Thanks for the info. Looks like I have some more reading to do before I decide to do some more tinkering.
 
To answer your 2nd question, with the CPU facing you the cores would be setup like this:

0 1
2 3

Hope this helps. :)
 
For those that don't know it yet, most core temperature programs have been misreporting core temperatures on Quad core processors. When you're Quad core temps are being reported high, low, high, low that is usually a pretty good sign of this problem. The reporting of the two center cores is likely swapped. CoreTemp is wrong, Everest is wrong and RealTemp 2.60 gets this wrong.

I'm hoping the present beta release of RealTemp 2.69.1 is finally getting this right but I need some more beta testers with confused Quads.

You can download the latest beta here:
http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

Unzip and copy the new RealTemp.exe into your present RealTemp folder.

This swapping of cores problem has been around since the first Quad was released. It can happen at any time after a reboot. My Q6600 was fine for 3 months before it swapped while some Quads swap reported core position on a daily basis.

It's very normal for the two sets of Dual Cores within a Quad to run at a different temperature. I think it is due to IHS to core contact. Let's post some data so we can get to the bottom of this problem once and for all. Too much time has been wasted remounting heatsinks trying to fix this issue when that usually doesn't help any.

AceGoober: This picture shows the correct physical core order.
http://www.neoseeker.com/Articles/Hardware/Reviews/core2quad_qx6700/12.html
 
Just tested the beta version of RealTemp and it reads the same as CoreTemp. I did notice that the core reading arrangement is different - RealTemp being 0231 and CoreTemp being 0123.

Thanks for the pic of the core arrangement. Definitely clarifies a few things for me. I am now under the assumption that I need to do a little more lapping as I believe that the IHS is a little high in the center yet. I remounted my block and my temps are now cooler in the center cores but this is before AS5 break in so I'll see what happens. I don't want to have to remount the block multiple times so once more, and that's it - I'm leaving it alone!
 
Here's a good test. Run Prime95 and go into the Task Manager and use Set Affinity... to force it to run only on CPU 0. On 65nm Quads this test clearly shows where your core 1 data is coming from. It should be coming from the adjacent core but if RealTemp is showing 0231 then CoreTemp will be reporting your core1 data as your last core which is wrong.
 
Ok , here's the conclusion I've came to since I did a little more research. - When installing the nozzle kit on the Fusion , there is a gasket that is placed in between the top and the midplate. From what I've read , this creates a slight bow effect similar to the Apogee GTX bow o-ring. When I installed the cooling on my machine , I lapped both the block and the chip to 2000grit. This obviously didn't change the bow as I removed the base plate of the block to do the lapping. As soon as I reassembled it , the bow was back thus creating a better contact in the middle of the chip. I think that this is the main contributing factor to having cooler cores in the center.

Maybe I'm just grasping for strings but it does seem very logical. I reseated my block multiple times and the effect was the same each time(when the pressure was evenly distributed across the chip).

Opinions on this theory?
 
I've got an Apogee bowed and HAD a Q6600 that had cores that differed by 5-7C and no amount of fiddling with the springs (hell I bowed the Apogee because this shit was driving me crazy AND now own a Q9450) changed a damn thing. It's just a deal with the quads. If you ask 4 quad core owners which core is the coolest you might get 4 different answers and if you ask them what the difference is between the coolest and the hotest you will get anywhere from 5-8C.

Don't loose any sleep over this.
 
Pretty much came to the conclusion that there is no uniformity what so ever. I'm just going to be happy with my low temps whether uneven or not. Thanks all that replied and gave experience/opinions.
 
You must log in or register to reply here.
Back
Top