[MaJoRoesch]

I'm working on overclocking my 3570K, and uh, every program gives me something different. I've tried PC Probe II (this is an asus board), speedfan, and realtemp. They all have different temperatures! How do I know which one is the actual temperature?

 

[UltraVolta425]

I work with Core Temp, I trust it.
For my GPU I use the ASUS GPU Tweak tool, it monitors it nicely.

 

[Recursion]

Open Hardware Monitor or read the temps from the ec? R&W everywhere can read temps from the ec. It's good to know when you want to make your own monitor.

 

[Agoriaz]

I trust CoreTemp and Realtemp since they show the same readings for me. MSI Afterburner for GPU monitoring and tweaking.
The reason for the different temperatures may be found in the individual software's settings. Check your settings and if there's an option for showing either Minimum, Average or Maximum temperature, go with either average or maximum. And make sure that it's set for the same in each program.
If there's still conflicting readings, I'd say go with either CoreTemp or Realtemp.

PC Probe is probably configured to show max reading, this is a guess though, based on AXTU (came with my ASRock mobo) which shows a higher reading than coretemp and realtemp.

 

[MaJoRoesch]

PC probe actually shows the lowest reading, to the point of sounding too good to be true. Realtemp shows the highest, BY FAR. Way way way way higher than PC probe. And speedfan is more or less in the middle. And we aren't talking small differences folks. While overclocked and running prime95 in my stress test, there was a 20C between Speedfan and Realtemp while under load. The difference from PC Probe to speedfan was 25C. This is total insanity. How is this even possible? There are only so many sensors on that CPU. Where are they getting this stuff from!
Quote:

The reason for the different temperatures may be found in the individual software's settings. Check your settings and if there's an option for showing either Minimum, Average or Maximum temperature, go with either average or maximum. And make sure that it's set for the same in each program.

I'm pretty sure that doesn't matter. I ran prime 95 for a half hour. No amount of variance from "average" versus "maximum" would account for these variances after a half hour running full blast.

 

[Coolwaters]

i use HWinfo64 just because its convenient. (it has everything)

also very accurate.

 

[Agoriaz]

That is insane. Maybe ( a lot of maybe's ) the software still needs some fine tuning on the developer's part to accomodate Ivy Bridge. It does however sound like speedfan is the unreliable guy in the gang.