Dual core temperatures

C47

I have a dual core cpu, and i get three readings from nxsensor; processor,core A, core B
when overclocking should i consider the processor reading?
during encoding these are the results
(proc,core A, core B)
PC Wizard 2007 : 49,50,55
nextsensor : 40.49,55
coretemeter : --.56.48
#1 : notice that nxsensor, pc wizard readings for both cores are both reversed when compared to coretemeter
#2 : proc temp in pc wizard is higher than nxsensor (either way it is lower that both cores)
so what reading does the processor correspond to (the processor die?)....and what significance does it have if overclocking

-----
another issue is that when idle there is a 10C difference between the two core....can someone explain why this is the case

Da_Jewish_Fonzy

I can answer one question, sort of, its normal for the two cores to be around 7 degrees different in temp with X2's from my expirience and also the expirience of others I've encountered on various forums, so don't worry.

As far as what temp to worry about, just pay attention to the core temps, most people follow the rule of thumb of no higher than 55 degrees core temp on X2's give or take, 57 or even 60 on one core won't kill your cpu but its best for longevity to keep them 55ish or under.

A good app that I like for monitoring temps is coretemp:

Core Temp

Its fairly popular and I've ccompared it to Everest Ultimate and it seems to also be pretty accurate or inline with the same temps.

PangingJr

good to read a very good comment and one that does not scare me away like that from you Da_Jewish_Hornet.
i just though i should tell you.

as for the op's post, what is your motherboard and processor type?

H3X4D3C1M4L

Two cores shouldn't be that far apart since they're located directly on the same chip... I'd use Core Temp. I've found it to be a pretty reliable program with minimal system resource impact

ChaosMinionX

Could be that far apart due to a number of things.... improper TIM application, uneven pressure across the IHS and heatsink base due to concave,convex surfaces on either IHS or base, or a misread by a faulty program.

H3X4D3C1M4L

Yeah but 10 degrees? It hardly seems likely even if all that is true (except faulty program)

ChaosMinionX

Its very possible actually Most quadcores average 5C difference across all 4 cores. 2-3 might be the same then 1 other is 5C higher or vice versa.

mkk

The sensors themselves aren't too accurate. There can be a large difference between the readout from the two cores simply because the sensors in each CPU deviate more or less in either direction from the actual temperature.

CDsDontBurn

what about TAT, the Intel Thermal Analasys Tool? whatever happened to that?

Lelisevis

My Quadcore seem to run within 1-2C difference across the cores at 100% load, though when I initially put it together I was getting upto 6c variation as the thermal paste hadnt spread very well.

ChaosMinionX

Thats a warm little B3 you got there

Lelisevis

Heh weathers pretty warm here at the moment and it's at 1.35v, never tops 60c though so plenty of playing room left. Oh and that AS5 was applied fresh that day so hasn't really bedded in. I'll give it a couple of weeks and bump it to 1.4v and 3ghz. Dont forget I got two 8800GTS cards in their and a stupid Asus chipset system which gets so hot I can't touch it.

PangingJr

nah i think the processor temps looks pretty okay to me, specially considering the system was in a warm environment.

Lelisevis

What a difference a few days of stress testing and burning in can make. Room temps still similar but im now idling at 36-37C. One degree difference on core1 is NWN2 updating. On top of that I have 3 messenger windows and 4 websites on the taskbar so it's not strictly idling.
Just goes to show that correctly applying thermal paste can make all the difference.

PangingJr

that Core temp also shows you need to burn in the processor some more (j/k).
actually, the temperature looks pretty good.



and i think 1-2 and up to 5 degrees celsius different between or among each core during system "idling" should not be much concerned to user.
however, when the system processor in "full load", the difference between each CPU core's should not be, at all time, more than 2°C.

and besides the CPU cooler, computer case air flow, (and just you might add the thermal compound to this), as you can notice, the room temps can/will make the big different on the system's and processor's temps. different room temps, different system temps.