i5 760 - 3rd core 10c cooler under load ?

MoodyB

MoodyB

MoodyB

Soldato
Joined
12 Jan 2003
Posts
3,749
Location
Motherwell, Scotland
Hey all,

Just received my i5 760 / Asus P7P55D and I'm currently trying to get it stable at 3.8ghz. Only concern I have is that in IBT or Prime, the 3rd core runs 10c cooler than the others at full load, and is 5-7c hotter at idle.

Only reason I'm asking is I haven't noticed as large a difference between core temps in various screenshots I've seen of 750 / 760 overclocks as I get here.

Think it's just a coincidence that it's this 3rd core that fails first in Prime if I don't have my bios settings correct, or do you think the chip is suspect ?
 
It could be a dodgy temperature sensor?
Or uneven spread of Thermal Paste.
Or as you say the core could be a bit dodgy, was it purchased second hand?
If not i would speak to the shop owners and see what they suggest.
 
It was bought brand new on Friday.

Think the thermal paste would be even. Used the 'half grain of rice' method in the centre of the heat spreader, as I noticed the stock intel heatsink had a circular base.

I use a Zalman Performa to cool it. It coped with an [email protected], so think it'll do fine with an i5.
 
Last edited:
Some variation is normal.... Mine has a core which runs cooler. 10 degrees sounds like quite a bit though. When you say the third core, do you mean the one labelled "core3" (ie, the 4th one, as the first one is labelled core0)? Because if it's "core3", that one is at the edge, so it could be uneven TIM. If the HSF was tilting one way even slightly when you screwed it down, it could have pushed more paste over one side.

Don't use IBT for heat testing, IBT uses linpack which doesn't give a very even workload. The work comes in peaks and troughs which is good for forcing errors (testing stability) but not as good for checking temperatures. If you look at the graphs after running linpack for a few hours, the temperatures can peak very high, but also drop down quite low at times as well.

Run prime with small fft's and see if that brings the temp up on core3. That test gives a more even workload, so it should give you a more even spread of temperatures across the cores (although the max temp might not peak as high as other tests).

Reapply the paste, and try small fft's. If the difference between cores ends up around 5 degrees, that amount of variation is perfectly normal.
 
It's core '2', going by the first one being '0'. I've been using Prime with Blend along with IBT to test out the cpu. I tried small fft's and it still showed 8-9c cooler on core '2'. It's slightly lower now because I lowered my vcore from yesterday when I was trying to get 4ghz stable on my i5 760.

I'm going to redo the TIM application as I only had AS Ceramiq at hand when my cpu arrived, so going to try some Gelid Extreme when it shows up, which will hopefully be tomorrow.

I'll post again on how it goes. Thanks for the advice.
 
3.2ghz on an i5 / P55 motherboard ?

If so, you should be able to get that at stock voltages, and probably reach 3.4-3.6 before you need to increase anything.
 
hi, i have the i5 760 oc'd to 3.8 stable. my 3rd core is also 8-10 dergrres lower under load but all cores are equal on idle. it dosnt really bother me tbh, but i guess your being higher on idle could be seen as suspect.
 
^^ if you can be bothered, then try redoing the TIM.

I redid it here using Gelid Xtreme, and the temps on core 2 have settled down to be inline with the others.
 
You must log in or register to reply here.
Back
Top Bottom