A Question Regarding Ivy Bridge Temp Sensors

[Wastedslayer]

I recently installed a new H110 in my system, and Ive noticed that while I run IntelBurnTest RealTemp seems to want to report two separate temps. It will float at ~60c and hold steady, then will jump to ~80c and hold steady and then back to ~60c and so on and so forth.

Each of these holds steady for 4-7s and then switches.

So which is the right temp, or what could be causing this?

 

[sneekypeet]

both are correct. IBT does not make the CPU run at its max the whole time of the testing.

 

[Wastedslayer]

both are correct. IBT does not make the CPU run at its max the whole time of the testing.

Good to know! Seems like I'm right in assuming that Ill have to de-lid this chip to help get it to more comfortable temps.

 

[EarthDog]

This delidding thing has gotten out of hand...

You can go to 90C while stress testing (100C is where it starts to throttle).

IBT is also MASSIVE overkill for stress testing IMO. THere is nothing you will do with a 'lesser' application, like Prime95 for example, that will even reach those temperatures, none the less an application like IBT. So I stick with AIDA64 testing and Prime95.

 

[sneekypeet]

When configured the same, I get the same max temps in IBT as I do in Prime.

As for delidding, what speed and voltage are we talking about here for that temperature range?

 

[EarthDog]

I dont touch any settings in P95 outside of Small FFT and Blend.

Crap, good points sneeky... Im hoping that is around 4.5Ghz, LOL!

 

[sneekypeet]

I dont touch any settings in P95 outside of Small FFT and Blend.

Crap, good points sneeky... Im hoping that is around 4.5Ghz, LOL!

well I run IBT and Prime with 3GB of ram being tested...I do believe IMC has a lot to do with the heat load

 

[Wastedslayer]

Im currently running 4.7Ghz at 1.295v

 

[EarthDog]

Id say that is about right then for that cooler... what are your ambient temps?

well I run IBT and Prime with 3GB of ram being tested...I do believe IMC has a lot to do with the heat load

Perhaps, but if that is the case, why does blend (which tests memory too) run a bit cooler than small fft in P95 (default settings - still uses a lot more ram than small fft)? There is more to it than that maybe?

 

[sneekypeet]

Perhaps, but if that is the case, why does blend (which tests memory too) run a bit cooler than small fft in P95 (default settings - still uses a lot more ram than small fft)? There is more to it than that maybe?

Don't really have an explanation currently. I will look into this though, that is a good point. I guess I am naive to the stock runs temperatures, because I was hunting to see if IBT was way out there, but with them both set similar in memory tested, I found they were within a degree or even tied in results on my 2600K.

@ OP, for 4.7ghz on that cooler, you are doing what I would expect. You could use better fans on that cooler, but I would guess you are going to only get five degrees of wiggle room at the expense of a lot of noise. The Krakens are basically the same units, and with just a fan change they do better then the Corsair versions, but again its just a few degrees.

Also delidding the proc is going to cause you issues with all of these coolers unless you are planning just to replace the TIM.

 

[Wastedslayer]

I should have mentioned I have Cougar fans on the Rad. As long as the temps are where they should be then I'm happy