Strange 3570k temps...

PowerPie5000

PowerPie5000

PowerPie5000

Associate
Joined
7 Sep 2012
Posts
957
Location
Lancashire, England
I've just received my 3570k from OCUK and the temps i'm getting for each core are quite uneven! I've reapplied Arctic MX-4 thermal paste and have reseated the cooler a couple of times but get the same results...

These are the idle temps i'm getting at the moment at stock speeds:

Core 0 - 22c
Core 1 - 27c
Core 2 - 33c
Core 3 - 30c

I got these temps using both 'Core Temp' and 'Real Temp'.... HWMonitor shows slightly higher figures (but the same large temp gap between cores). I did have an i5 2500 installed which only had a slight variation of 1~2 degrees between cores (it would idle at around 28~30c using the same cooler).

Has anyone else experienced this?
 
From what I've seen, it's not uncommon to see a wide spread of temperatures. Mine is similar but not quite so much as you have - about 5C from lowest to highest. On mine core0 is much closer to the other readings than yours is.
 
Maybe the surface of my CPU is slightly uneven? I've never seen such uneven temps (Theres more than 10c difference between core 0 and core 2 :eek:).... I'm going to stress it a bit and see how much the temps vary under load.
 
Yeah, my i7 920 had that, but with Core 0, 2 @ 31c, Core 1, 3 @ 35c.

I'd go with the slightly uneven cooler. So try tweaking it to get it totally flat, and it may sort out the problem.
 
8-10c difference between cores is normal for IvyBridge

Idle temps are never accurate and can often show values below ambient. I'm guessing your room is at or around 22c right now, so Core 0 is clearly wrong anyway.
 
spixelspixel said:
Run prime95 and see if they even out.
Click to expand...

I ran prime95 for about 30 mins after my last post... The temps were:

Core 0 - 56c
Core 1 - 58c
Core 2 - 62c
Core 3 - 60c

The difference between core temps is a bit better under load :). I'm using an Arctic Freezer Pro 7 Rev 2 cooler (cheap, but still ok ;)). My i5 2500 had slightly lower temps using the same cooler though.

CPU-Z is reading a core speed of 3607.5 Mhz whilst running Prime95 on all 4 cores. It's also reporting the CPU voltage as 1.176V (i leave all this stuff on auto in the bios and don't overclock).
 
Last edited:
Unless one of the cores shows a difference more than marginal e.g 3cores at around 20-30 and final core at 60 , then I wouldn't worry too much about it.

Nothing is perfect .
 
I reckon my 3570k has a dodgy sensor at idle. Cores reading on coretemp at 28-29-27-19 sometimes. Considering that my ambient room temp is 22c, 19c isnt possible with the cooling i have. Every intel quad ive owned has had a bit of core variance, including a q9550 that had a stuck sensor at idle. Even after lapping, my old q6600 still had the same variation as before.
 
The sensors only role are to prevent the chip from overheating, so they aren't calibrated to give accurate readings below a certain level. If the cpu is running below lets say 30c there is no need to be accurate, since there is no action required (fan speed increase/thermal control/thermal shutdown).
 
My 3570k temps are similar, core#2 can be 7-10 degrees cooler than the others.

Idle temps:

idle-temps.jpg


Prime95 temps:

Untitled.jpg
 
I'll just leave my CPU as it is... Seems a common thing for Ivy Bridge CPU's. I know it's not the cooler as you can't really go wrong installing an Arctic Freezer 7 Pro Rev 2 (securely attach the push-pin bracket, then i apply a thin even layer of Arctic MX-4 on the CPU and lastly bolt the cooler to the bracket using the 2 screws... Fool proof installation ;)).
 
Mine are the same dont worry. Overclocked to 4.4ghz with Max Temps: Core 1) 62 2)74 3)71 4)67. So some variation! Temps should improve the more i use it, but dtill have that variation
 
I've only just discovered that Intel cheaped out on the internal thermal paste with Ivy chips (between the CPU and heat spreader). Sandy Bridge apparently uses fluxless solder which will last a lifetime and conducts heat better... Ivy Bridge on the other hand uses some cheap white goo that will no doubt evaporate, dry up or just get worse in some way over time :rolleyes:

... The variation in core temps with my 3570K (and others) is probably due to the internal paste not being applied evenly.
 
I had the same thing with my newly installed 3770k over the weekend, pretty much the same temps and differences between. I even re-seated it with new paste and got the same results.

Idle: 30~
Prime95: 60~
Stock clocks
 
Last edited:
Maacck said:
I had the same thing with my newly installed 3770k over the weekend, pretty much the same temps and differences between. I even re-seated it with new paste and got the same results.
Click to expand...

I don't think any amount of re-seating will solve this one... I've done it a few times with the exact same results. Must be down to dodgy sensors or the crappy paste Intel applied under the heat spreader. I can't believe this 22nm chip runs slightly hotter than my 32nm i5 2500 (which is now in my girlfriends new build) at stock speeds :(.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom