Weird temp readings.

[Extremez]

I just upgraded my computer yesterday with an E8400 and EVGA 780i mobo (along with other parts of course). So I got it all up and running and figured I'd check out my temps (since I added 2 more 120mm fans to my case). HWMonitor was showing me 72C on both cores idle with the proc underclocked to 2GHz (bios setting for idle I haven't changed). I got a temp gun out and started shooting anything I could get at the base of the heatsink since there is no way to get the proc itself and the max I got was 30C. I'm not sure what sensor it uses for the cpu temp in the bios but that also read 30C when I rebooted. I let my computer sit for half an hour off (also overnight later with same results), started up HWMonitor as soon as I could and it was at 70C cold. In HWMonitor I noticed there is a System temp, CPU temp and a temp for both the cores. The system is reading -64C and doesn't move so I don't know if that means I lack a sensor? The CPU temp reads 26 idle and 42 full load, and the cores are at 72C 77C max. So to make a long story short, does this mean my actual core sensors are wack, and if so would it be smart to RMA it (ie if I hit 50C actual will it think it's 100C an throttle down?)

PS I didn't run prime too long, but I did play games for hours straight with the same temp results.

 

[metloaf]

I had a similar issue with my 3110 when I first got it. These CPUs do have an issue with the temp sensors as far as reading idle. Mine was running in the 70s initially and I had to reseat the water block and that fixed the issue. Now I am running in the 20s idle and 40s under load. I don't know what kind of cooling you are using but try to reseat your cooling block/heatsink again and go from there.

 

[metloaf]

Oh I almost forgot, welcome to the forums.

 

[Extremez]

I've read a lot about the idle temp problems on these, but never found anything this far off, and if it was a seating problem wouldn't the temps still start off low and work their way up?

 

[blazed]

First you should use something known to work well, RealTemp or use CoreTemp and subtract 10C (due o 95 vs 105C Tjunction max) from what it gives you.

But I agree with metloaf, sounds like your heatsink isn't installed properly. Your heatgun can't measure the core temps under the heatspreader, and your heatsink won't warm up much if its not making proper contact.

 

[Extremez]

I have used all of those programs and they all read the same (with the 10C taken into consideration) I do not see how it could be a bad seat due to the following reasons.

I did a cold boot and it read 70C the instant I got it up. If indeed it was a bad seat and got to 70C that fast it would reach 100+ shortly thereafter (which it does not).
On a bad seat the cold boot temps still remain very close to the ambient temp for a short period but if you leave it just idling the temps will keep climbing and climbing, and under max load the temps will shoot up considerably (both do not happen).
Another factor but not very accurate is the actual temperature of my room. My last proc ran pretty hot (P4 3.6GHz Prescott) and since my room is small after a long night of gaming it would get pretty hot in here. It is noticeably cooler in my room after changing over which again leads me to believe that the proc is not running hot.

 

[Technophobe]

The CPU temp reads 26 idle and 42 full load

I think this could be an indicator if your HSF is mounted properly or not. I'd assume your idle CPU case temp would be a lot higher if you had seating problems.

Does BIOS affect the core temp readings (obviously not the actual temp)? I'd update the BIOS anyway if it needs it.

One way I can see having 70c+ core and only 26 CPU case temps is if the actual CPU heat spreader (top) isn't making good thermal contact with the cores.

Anyway, I'd reseat the HSF, use good TIM just to be sure.

 

[Extremez]

I just reseated and have the exact same temp readings. I guess I'll see if I can update my bios even though I'm not sure if that has anything to do with the core temp readings.