TAT or coretemp, which is considered more accurate?

[pieman109]

As title. I'm generally getting higher temps indicated in TAT than the temps indicated in coretemp, not massively different though:

TAT.....35c/31c idle
coretemp.....30c/28c idle

The load temps when running orthos are both within 1c of each other

TAT.....44c/43c load
coretemp....43c/42c load

This is on an e6600 @3.2ghz, 1.4vcore

Which should I accept as being the most accurate?

 

[Tooks]

Well, generally Coretemp is regarded as the most accurate, but in theory you'd think that an Intel tool should be able to read and extrapolate the signals from the in chip thermal sensor as well/better than anybody else? But, Coretemp is regularly updated to take account of new CPU's and I don't think TAT is anymore, as it was only ever meant for internal use by Intel engineers I understand.

To be fair, they are both close enough at load to accept either as being relatively accurate.

Which is the main point really - that whatever app you use to measure your temps, they are only ever relative to your own system, and not for comparing systems. As long as you only use them to monitor your own system and to see how changes to it affect temps etc, then either would be fine.

Personally, I go with Coretemp. TAT won't even run on my current setup with a Jan 2008 E2180.

 

[wolvers]

Those temp readings aren't very far apart anyway!

Seems like you're running pretty cool too chap.

I trust TAT more than coretemp for my L2 E2160. I'm not convinced by the Tjunction in coretemp for it, although TAT was originally meant for mobile cpus so there's no guarantee that it's right either.