Same cpu-different boards. Is there a true effect on temps?

[Texun]

This is not a cooling question or a question about sensor accuracy, but a question about the possibility of one board actually causing the same cpu to run hotter or cooler that another board.

There are countless threads noting the difference in temp readings from one board or BIOS to another. I can understand some minor differences based on sensor calibration, type, and location on the board but my question is this:

Temp sensor variances aside, can one board actually cause the same cpu to run hotter or cooler than another board? Example: Board A shows a cpu temp of 48C but when the same setup is moved to board B the temp readings show a change of 12C. Sure, part of it is a calibration problem, but do some boards actually run the cpu hotter than others?

My ASUS board showed a temp of 51C and the identical setup in my Abit shows 37C. Sure there are sensor variances and I don't care about those as long as I know the difference, but I would almost swear that the Abit feels cooler when I give it the old "finger test." Even after a couple of hours of hard use the area around the socket and HS base are just slightly warmer than the rest of the board or cards (except for the AGP). Maybe it's just my imagination but it seems like the Abit board actually runs cooler. The case temps are also about 3C cooler now and I used the same parts and thermal compound. I would have thought the difference was all sensor error but maybe the board does make a difference. Hmmm.....

 

[floccus]

Possibility. With different implemtations of the northbridge you could possibly get a board that runns a little hotter. I could really see this as the case if you're running a fast FSB and the data rate between the northbridge and memory are high. Because different manufaturers have different channels between the northbridge and memory, if one is running at a higher frequency then it seems logical that it would run hotter.

But I still think there is a big difference based on thermal diode implementation. The only real way to check your claim would be to user a laser thermometer and check the temps of the cpu and northbridge. That would yield the most accurate result and then we could see if there is a difference between boards.

 

[beatle]

Maybe not 12C difference, but voltage can vary quite a bit. My NF7-S reports 1.65-1.66 vcore when it's set to 1.7v in bios. Some boards overvolt. More voltage = more heat.

 

[Trader05]

I just upgraded my board,asus p4b266c to abit bh7. Same voltages and everything, cpu temp went from 34 idle (asus) to 48 (abit), why are these numbers jumping?!??!

 

[Texun]

Originally posted by: Trader05
I just upgraded my board,asus p4b266c to abit bh7. Same voltages and everything, cpu temp went from 34 idle (asus) to 48 (abit), why are these numbers jumping?!??!

I wish I knew. Mine went the other way. The Abit that I have now has never gone above 40C while the ASUS did not want to drop below 50C. I know the calibration has a lot to do with it, but I don't think it is too much of the end user to ask for them to be closer than this! I see a good 12-15C difference with the same hardware on different boards.

I also have the P4B266C with 1.6 Northwoood. I don't remember the temps at the moment but they were never high. I would have remembered that.

 

[al799cmt8]

My epox 8k3a+ temp is 57 C while epox 8rda is 47 c. Using my fingers I can notice the HS is much hotter with 8k3a+. The HS in 8rda is only warm!