Different core temps / Oscillating clock speeds

horrorxpunk

horrorxpunk

horrorxpunk

Associate
Joined
5 Jul 2012
Posts
384
Location
Sheffield, UK
I searched all over the forum for a thread about this but I am having two problems or rather concerns involving two possibly related stats.

yWwiiKD.png


Firstly my there is some distance between my core temps, the room temperature is currently 16 degrees; they used to be more even but have since changed while I have been overclocking heavily I understand this can be normal, but core #3 was reporting temps below room temp and core #2 always runs hotter than all the rest.

The second problem I am having is with oscillating clock speeds, I have switched off speedstep and C1E but my clock speeds still oscillate along with my voltages 3899.78~ - 3900.15~ Ghz 1.168v - 1.176 volts I know these are minor and I can get overclocks as high as 5.0 Ghz I am looking to push this further with a new kit I am wondering if this could cause instability or not or if there are other settings in the bios I could change to stabilise the core speeds I am using an Asus Maximus V Formula.
 
Disabled EIST?

The temp on #3 is most certainly wrong, so its probably down to one of these things....

The sensor is 'stuck' or faulty
HWMon isn't reading the chip correctly, try realtemp/coretemp

Do the temps even out under load?
 
All programs match, I have foudn HWMon to be pretty good so far.

2v9necn.png


I'll never get an RMA for that chip, I've overclocked it at 5.0Ghz with 1.506 volts. The sensor is located on the chip right?
 
My 3DMark 11 physics scores are lower comparing a 4.9Ghz with a 5.0 Ghz bench.

Old result for 4.9Ghz

http://www.3dmark.com/3dm11/6102312

Recent result for 5.0Ghz

http://www.3dmark.com/3dm11/6245464

Is there a way to even these temperatures out using a torture test?

If the core has borked I will deffinately be replacing it with the 4770k it has a lot of performance benefits for me due to the new iGPU I run LucidLogix VirtuMVP which combines the power of my iGPU with my discrete card it gives me around +50% performance improvement with the 3770k.

EIST is deffinately disabled as is C1E.
 
Last edited:
Looks like either a duff sensor, motherboard reporting error or it could even be the CPU block is on slightly too tight.

The temps line up fine on your H70?
 
I'm not using the H70 anymore, but no I was having the same problem with the H70 before I don't think it's the CPU block, this is my new rig.

Co4WKUP.jpg
 
In which case I'd point the finger of blame to either a borked sensor or the motherboard reporting.

You on the latest bios for the motherboard?
 
Pretty sure it's 1603, it's a shame I can't check temps in the bios, going to have to get a thermometer straight down on it hehe. I'll just go by my hottest temperatures even if I can't trust them. But it doesn't explain the drop in physics scores, core #3 may have slipped away during the night. I often run prime95 tests though and all the cores match up okay. Let me give it a 10 minute blast now, Core #3 will normally stick 5 degrees behind the hottest core minimum.
 
Intel's core temperature sensors are only designed and calibrated to control thermal throttling and thermal shut down. They are not even close to 100% accurate so you should not expect to see your temperatures line up perfectly, especially when the CPU is idle and the temperature is a long ways away from the calibration point.

Accuate to +/- 5C at the calibration point is typical so your sensors are well within spec.

You are best off to keep an eye on the Package temperature when lightly loaded. This combines data from all of the individual core temperature sensors and helps reduce the amount of error.

RealTemp T|I Edition
http://www.overclock.net/t/1330144/realtemp-t-i-edition
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom