odd temperatures

[matty-hodgson]

yeah, i suppose it is the only way i'll learn.. and it'll all help me when i'm older because i want to take up a job in IT.. good money, and i find it interesting. only bad thing is, i didn't take IT at GCSE so next year in sixth form i'm taking 2 courses in IT, 1 will take a year and i do the GCSE stuff in that lesson and the other 1 i do for both years and that's the proper A-level course..

so i've to let my PC warm up and then remove the heatsink and get the grease off while it's still warm?

how long does it take to cool?
i'm guessing it'll cool quite quickly with it being such a good thermal conductor, also losing its heat quicker becuase of this..

is that 1 of these huge heatsinks that are like a full 3 kilos of metal hanging at 90' from your motherboard?

they look very scary! and i wouldn't be able to trust 1..

[matty-hodgson]

it came with the original Intel thermal paste, which i doubt is good.. and i also rubbed a bit of it off to check if mine was the G0 or GO stepping, i dunno which it is..
and i think that's the culprit for the difference in temperatures between the 2 sets of cores..

what is the max that you think a Q6600 should get to?

i'll keep using core-temp then

so the temperatures your CPu is giving out depends on what task it is performing? am i understand that correctly.. if the L2 is being used more or something?

[Fraz]

Gah - you ask questions faster than I can answer them!

Use CPU-Z to find out what stepping your CPU is. And it's G-zero, not G-oh.

A Q6600 G0 core can tolerate up to 100 degrees centigrade for short periods, apparently. (I believe this is sort of the meaning of the "Junction" temp that you can see listed in CoreTemp). However, the specs say that the "case" temp (normally ~10-15 degs lower than a core temp) should not in general exceed 70 degrees. So I guess your cores should not exceed ~80 degrees on a long-term basis.

The answer to your last question in complex... Yes - the temp of a core at 100% load depends heavily on what code that core is executing. If you engineer your code very very badly so your CPU is continuously stalled with, say, memory requests to the main system RAM, then it won't get very hot. It is these stalled cycles that Intel's HyperThreading technology on the Pentium 4 processors were designed to use... whilst the main thread is stalled, a second thread can make use of the CPU for a few cycles. Quite clever really, but not a feature of a Q6600.

[Fraz]

yeah, i suppose it is the only way i'll learn.. and it'll all help me when i'm older because i want to take up a job in IT.. good money, and i find it interesting. only bad thing is, i didn't take IT at GCSE so next year in sixth form i'm taking 2 courses in IT, 1 will take a year and i do the GCSE stuff in that lesson and the other 1 i do for both years and that's the proper A-level course..

Why not do Physics...

Less well-paid, but you get to work on stuff like this, which - let's face it - is a damn site more interesting than IT.

[matty-hodgson]

i'm sorry.. do you want me to stop asking questions?
i think i'm finished asking them anyway soo..

apart from these last few...

WHAT THE BLOODY HELL IS THAT IN THE PHOTO?!
and.. so anything below 70'C at load is acceptable, but if you can get lower, get lower..

and.. do you use a big, hefty block of metal as your heatsink.. i dunno if i could trust 1, do you need a strong motherboard for 1 of those?
it's only PCB.. :S

[Fraz]

i'm sorry.. do you want me to stop asking questions?
i think i'm finished asking them anyway soo..

apart from these last few...

WHAT THE BLOODY HELL IS THAT IN THE PHOTO?!

Compact Muon Solenoid - Wikipedia, the free encyclopedia

and.. so anything below 70'C at load is acceptable, but if you can get lower, get lower..

yup, that sounds about right to me. Best be on the safe side.

and.. do you use a big, hefty block of metal as your heatsink.. i dunno if i could trust 1, do you need a strong motherboard for 1 of those?
it's only PCB.. :S

Not especially. My heatsink weighs nearly a kilo, or something stupid, and it's not the heaviest out there.... As long as you don't move your PC around much, it's fine. If you do move it around, do it gently, or hold the case such that the motherboard is horizontal

[matty-hodgson]

i'd never, under any circumstances hang a kilo of metal at a 90' angle off my motherboard.
it's total insanity!
i don't really know the formula for working out the turning force of the end of that, but it must be a lot!
the Intel 1 doesn't weigh much and doesn't hang off much at all.. a heatsink i trust

and thanks for all the help aswell pal, i think i'll add you as a mate if that's okay

cheers bud

[matty-hodgson]

just applied arctic silver, dropped from 45'C idle (average of the 4) to only 32'C at idle..
i might even overclock it a bit now.
apparently the paste gets better the longer its been left in, obviously up until a certain point so.. ALL GOOD

thanks for all the help guys.
but the CPUs are still slightly different in temperature, only 3'C difference though.

[Thorsson]

Isopropyl Alcohol is a liquid that evaporates very quickly, this it does not hang around to cause a short like water would. Also it is inert, which means it doesn't react with metal, which tap water (with Fluoride in it, thus making it mildly acidic) could do on prolonged contact. It is thus the perfect cleaner for the CPU.

[matty-hodgson]

but what exactly do you need to clean it for?
that's what i don't understand.. surely giving it a good rub with some tissue paper will get it all off, right?

[Webby]

Tissue can only get rid of so much the IPA finishes the job to get you a clean surface for the reapplication of thermal paste. Oh and IPA is non-conductive it also comes pure unlike ethanol or acetone which tend to have more water in solution which is bad for computer components, this is more important than its speedy evaporation but that doesn't hurt!

As to tap water being slightly acidic due to fluoride its much more likely to be the H+ ions that are doing it