B3 Q6600 core temp differences

[mDust]

I was recently bugged that cores 2 and 3 (of 0,1,2,3) are always ~5C below 0 and 1. It's the same difference at idle as it is with all 4 cores at 100% load. I tried re-seating the water block to no avail. I tried restoring stock clocks and even underclocking. I then lapped the water block very nicely to a near perfect mirror finish...still no change. I'm thinking of lapping the processor now...
I believe this temperature difference has always been the case, but I've been trying to figure out why this has been happening. I suspected air bubbles trapped in the water block, and after jumping the PSU and rotating the block around with water flowing, some air did pop out. This lowered the overall temperature by about 1-2C but did not close the difference at all. With Google's help I've read a lot of questions about this but the responses were generally that this is normal and nothing can be done about it. So I'm thinking maybe it's a software issue: maybe Windows services and other programs seek to run on the first two cores before the last two. I don't know anything about how Windows goes about assigning work to a core...perhaps it's just not an even selection process and favors 0 and 1? I don't know if I can actually find out what is running on which specific core... In Windows Task Manager under the Performance tab, I watched the cores idling and 0 (when idling) always seems to have the least amount of usage...yet it's usually the hottest by 1C.
Anyone have any thoughts before I do something dangerous to my CPU?:D

[NightrainSrt4]

I don't have any real ideas on what you could do to fix it. But if it is bothering you, you can use realtemp to check your temps. If it is still showing up, you can adjust the temp readings of the cores (as the probes sometimes just don't match up even if the temps really do) so they are all relatively even.

[Spawn-Inc]

i had my cores like that on my q6600 G0 and after i lapped it cores 0,1,2 are the same and core 4 is 1C off.

but yes, thats a perfectly normal issue and nothing to worry about.

[crenn]

Same here with my Q6600 B3.

[mDust]

perfectly normal issue and nothing to worry about.I'm not worried, just inquisitive at this point. I think I'll lap the processor just for the experience. It will be the first time I've done a processor...so I'll think through the process carefully before I do anything regretful.
I did some uninteresting reading about how the temp sensors on the cores are analog and their signal has to be converted to digital...and how they aren't completely accurate in the first place and the conversion throws them off more...and how the temps any processor reports are estimates...and how some software reporting programs report the wrong temps despite what the CPU reports anyway...
...I've come to the conclusion that nobody has any idea what temperature their cores are running at...not even the chip manufacturers. :banana:

[mark22]

I don't have any real ideas on what you could do to fix it. But if it is bothering you, you can use realtemp to check your temps. If it is still showing up, you can adjust the temp readings of the cores (budget hosting (http://www.businesshostingprovider.com/budget-web-hosting.htm) as the probes sometimes just don't match up even if the temps really do) so they are all relatively even.

I think it could be that the HS is not seated properly. I'm not sure about the B3 stepping, but the G0 IHS is just a fraction of a mm shorter than older 775 chips (read my old D820)- meaning the CPU retaining bracket may interfere with the HS properly contacting the IHS.

:bunny::bunny:

[mDust]

I think it could be that the HS is not seated properly. I'm not sure about the B3 stepping, but the G0 IHS is just a fraction of a mm shorter than older 775 chips (read my old D820)- meaning the CPU retaining bracket may interfere with the HS properly contacting the IHS.

:bunny::bunny:
Thanks for the info. I've reseated it a dozen or so times, so it's as good as it's going to get. It's 6 months later now and nothing has changed...37, 34, 33, 33 *C under a light load.

[SXRguyinMA]

the Q6600 I just got from trace runs temps very similar to yours, with cores 3 and 4 running cooler

[slaveofconvention]

If you look at my recent Thermaltake Frio review, you'll see my Q6700 has minor temp variations between the different cores and I can fully assure you that was definitely mounted properly - it's just one of those things - you have 4 very similar but slightly imperfect temperature sensors monitoring 4 very similar but slightly imperfect cores - there's always going to be a little variation in the numbers.

[Kayin]

It could also be IHS deformation. Lapping would reduce those temps, and might pull them in line. For proof, let me link you an article that shows my old Q6600's TIM footprint...

http://www.thebestcasescenario.com/frontpage/?q=node/368

As you could see, the IHS was BOTH convex and concave. Not too healthy...

[mDust]

It could also be IHS deformation. Lapping would reduce those temps, and might pull them in line. For proof, let me link you an article that shows my old Q6600's TIM footprint...

http://www.thebestcasescenario.com/frontpage/?q=node/368

As you could see, the IHS was BOTH convex and concave. Not too healthy...

Yeah, that's what I thought, so I lapped both the heatsink and processor. It brought the temps down, but the gap remained. As it turned out, my IHS was concave with a convex bump in the middle as well. I had to sand it down to the copper to flatten it out...it was pretty bad.

[billygoat333]

I think it could be that the HS is not seated properly. I'm not sure about the B3 stepping, but the G0 IHS is just a fraction of a mm shorter than older 775 chips (read my old D820)- meaning the CPU retaining bracket may interfere with the HS properly contacting the IHS.

insert random unrelated link here (www.thebestcasescenario.com)
:bunny::bunny:



:D