Q9300 Temperatures

[Sectick]

Here is my setup:
Q9300 (Only running at 2.3 Ghz - Need a bios flash)
Zalman 9500A w/ Artic Silver 5 thermal paste.

It should be known I am new to overclocking, but I have read the sticky and understand the basics.

Here are my results:
http://i13.photobucket.com/alb...ctick/PC/CoreTemps.jpg

My questions:
1.) What is more accurate for my processor RealTemp or Core Temp?
.......(From my readings it sounds like Realtemp is more reliable for 45nm, any verification?)

2.) Why is there such a huge difference in core temperatures?
.......(I followed the directions for artic silver 5, made sure everything was cleaned properly and had no ooze out. It was applied exactly the way specified in the AS5 instuctions.)

I am getting ready to Overclock, but the large core temp readings made me take a pause. If the realtemp readings are accurate I think I can manage 3.0 Ghz even with the huge variance in core temperatures.

Thanks in advance!

 

[bryanW1995]

core temp and real temp are both accurate, they just don't agree on what tjmax is. If you look closely you'll notice that they're reporting 10c difference on all 4 cores. real temp thinks that tjmax is 95c, coretemp thinks that it's 105c. it doesn't matter, however, since they both accurately report distance to tjmax the same. just either use realtemp only or have coretemp display distance to tjmax and you'll be gtg.

well, other than the fact that you have too much delta in your core temps across the 4 cores. you need to re-seat your heatsink and reapply thermal paste. sorry, I know that it sucks, but you'll be glad you did later!

 

[Sectick]

Sorry this might be a completely noob questions, but by delta do you mean difference in temperatures?

Also, any tricks for applying AS5, because like I said we took great care to make sure we set it down square. So if I am going to do it again I'm not sure how I could do it any better.

I'm hesitant to pull that thing off again (I need to remove my whole MoBo) but if it needs to be done it needs to be done.

**Edit** I should also say those temps are running under load w/ Prime95

 

[bryanW1995]

yes, it sucks having to remove your mobo to take off the hsf, but yes, you need to reseat it. there are different theories on how to best apply the thermal paste. often you can have too much applied, so be careful to just apply enough to cover the ihs (integrated heat sink - this is what most people think of as the top of the cpu). try this: apply a pea-sized amount to the ihs, then squish it around with the hsf before re-affixing it. this is often the most effective method.

 

[OCGuy]

those temps look fine under load...

 

[Drsignguy]

Syzygies just posted a thread of what seems to be what your two core temps are showing. No answers at this point but I am sure they will come. I think this will help all of us.

 

[BTRY B 529th FA BN]

When he is testing for stability of the cpu, which is a better stability test, OCCT or Prime95?

 

[myocardia]

Originally posted by: BTRY B 529th FA BN
When he is testing for stability of the cpu, which is a better stability test, OCCT or Prime95?

Both, bot OCCT will find problems much faster than Prime95.

 

[BTRY B 529th FA BN]

Any particular program setting in both? I've heard you are suppose to set something in Prime to detect errors. What settings should us dual and quad core people use?

 

[Foxery]

Originally posted by: Sectick
Here are my results:
http://i13.photobucket.com/alb...ctick/PC/CoreTemps.jpg

My questions:
1.) What is more accurate for my processor RealTemp or Core Temp?
.......(From my readings it sounds like Realtemp is more reliable for 45nm, any verification?)
2.) Why is there such a huge difference in core temperatures?

For some reason, these programs measure temperature by subtracting from the chip's maximum rated tolerance. ("TJmax") CoreTemp believes this is 105C, and RealTemp believes it is 95C. Either way, what Bryan is trying to say is that as long as you stay quite a bit below the max, (25+ degrees) you are "safe."

Your cores shouldn't be quite this much different; I agree, definitely reseat the heasink and use less paste this time. I went through the same, and now they stay within 2C of each other at all times.

I am getting ready to Overclock, but the large core temp readings made me take a pause. If the realtemp readings are accurate I think I can manage 3.0 Ghz even with the huge variance in core temperatures.

Reach higher!

 

[ShadowFlareX]

Did you use Arctic Silver's method on how to apply the Arctic Silver 5 thermal paste onto your Quad Core? I did and worked best.

Arctic Silver Instruction

 

[bryanW1995]

intel has been very sneaky about giving info regarding tjmax on all core2 based cpus. The only people who know WHY they've refused to divulge this info are under nda. This forces all 3rd party temp monitoring software encoders to "guess" at tjmax if they want to report temps, which is why I don't even care what "temps" my coretemp/realtemp/speedfan/ir thermometer/abacus gives me.


for more info, here is unclewebb/kevin glynn's original post from XS (moved to his notes section on techpowerup). this is a MUST READ for anybody planning to OC a 45nm based cpu, and recommended for all other serious oc'ers.

http://www.techpowerup.com/realtemp/docs.php

 

[Sectick]

Just a quick update that you all might find interesting. I decided to try one more thing before I pulled the heatsink to make sure all variables were fixed before executing my last option.

I updated my motherboard bios (aBit IP35 Pro) from 14 to 16. Presto all temperatures are within 7 degrees.

Now my core readings are:

Idle I'm sitting at 41-42-39-35 and at load @ 3.1 Ghz 57-57-54-52

And cpu temp readings using CPUID" (Before it was freaking out)

Idle: 34
Load: 49

So, while my heatsink might still need reset I am happy with these results for now. Am I wrong in not resetting my heatsink right now? Like I said I am new to overclocking, but if I am understanding correctly all those temperatures are safe.

In the future, if I decide to push past 3.1 then I will reset it.

 

[Foxery]

BIOS updates are key

5-6C difference isn't stellar, but yes, overall this should be "safe." I agree that reseating is a good idea if/when you push higher. I've cut mine to a 1-2 degree variance.