Quick question about thermal paste and CPU temps

P

pdp76

Weaksauce
Joined
Jan 31, 2006
Messages
113
I recently switched from a PD 945 to a E6300 C2D. I'm using the same heatsink (Thermaltake TR2 M21), chassis and power supply. The only thing that changed is the motherboard and the thermal paste on the CPU. Anyway, when I had the 945 in, the temperature of the CPU was around 30C idling in windows. Now the E6300 is idling at around 47C in windows and I'm thinking this is a little high. First of all, I'm pretty sure the E6300 isn't supposed to be significantly hotter than the 945 right? 2nd of all, can the type of thermal paste make that big of a difference in CPU temperature? And in case you're wondering, I'm pretty sure the temperature readings are correct on both the old and new motherboard.

One last thing, I've read that when in the system BIOS, the temperature tends to be higher than when in windows (not sure of the exact reason). In the case with the C2D, the temperature read in the BIOS is about the same as what's read in windows, would this be a clue to why the C2D temp is pretty high?
 
I was using the stock heatsink/fan with the stock thermal paste and then took that off and added some Artic Silver. My temps basically stayed the same...if anything they went up a couple degrees. I idle around 45 and load is about 60 according to Easy Tune. When I use CoreTemp the temp is even higher for load....around 66C. Keep in mind that I am overclocking a e6300 to 2.8GHz. I don't know what the temps are stock. I don't think I ever ran stock. I just went straight to 400 x 7.
 
Lyquist said:
I was using the stock heatsink/fan with the stock thermal paste and then took that off and added some Artic Silver. My temps basically stayed the same...if anything they went up a couple degrees. I idle around 45 and load is about 60 according to Easy Tune. When I use CoreTemp the temp is even higher for load....around 66C. Keep in mind that I am overclocking a e6300 to 2.8GHz. I don't know what the temps are stock. I don't think I ever ran stock. I just went straight to 400 x 7.
Click to expand...
I'm not even overclocking the thing, just running it stock and yet, I'm getting similar temps as you are! Any other ideas what might be wrong?
 
There are three possibilities the way I see it...

1) Your hsf is not mounted correctly or you have too much or too little paste on.

2) You just got a chip that runs hot.

3) What you are using to get temps could be wrong???
 
pdp76 said:
I recently switched from a PD 945 to a E6300 C2D. I'm using the same heatsink (Thermaltake TR2 M21), chassis and power supply. The only thing that changed is the motherboard and the thermal paste on the CPU. Anyway, when I had the 945 in, the temperature of the CPU was around 30C idling in windows. Now the E6300 is idling at around 47C in windows and I'm thinking this is a little high. First of all, I'm pretty sure the E6300 isn't supposed to be significantly hotter than the 945 right? 2nd of all, can the type of thermal paste make that big of a difference in CPU temperature? And in case you're wondering, I'm pretty sure the temperature readings are correct on both the old and new motherboard.

One last thing, I've read that when in the system BIOS, the temperature tends to be higher than when in windows (not sure of the exact reason). In the case with the C2D, the temperature read in the BIOS is about the same as what's read in windows, would this be a clue to why the C2D temp is pretty high?
Click to expand...


What kind of overclock did you see with that pd945?
 
loxety said:
What kind of overclock did you see with that pd945?
Click to expand...
What exactly do you mean by that? I didn't overclock it, just ran it at the stock 3.4GHz speed.

And to respond to Lyquist's last post:
1) I removed and reseated the HSF and redid the paste twice and I got the temps down to where they are now, it used to idle at 55C (I used way too much paste) Either way, like you said, I don't think I'd get the temps down to 30C, even if I put the exact amount of paste and seated the HSF perfectly.

2) This is what I was kind of afraid of, that I just got a "lesser" CPU. If this is the case, should I try and exchange it or does it not really matter if it runs a bit hotter?

3) I guess it's possible the temps are still being reported incorrectly. Should I try putting a thermometer at the base of the HSF? =)
 
This is my question too about the difference in thermal paste.

because before i used artic silver in my p4 3.0C and was seeing REALLY high temps 46 degrees idle. and i applied and removed and cleaned it properly and i dunno i just had such a hard time getting the low temps. then i got fed up and i used some white generic thermal paste and i got my system to go down to about 31-33 degrees C.
 
I bet you need to lap that stuff. i.e. the heatspreader on the conroe is not very flat.
 
BillParrish said:
I bet you need to lap that stuff. i.e. the heat spreader on the Conroe is not very flat.
Click to expand...
I agree with you on that , i did mine and its running at 60c on load at 3.33ghz , will go down a fair bit in the next two weeks .

Warning- you will lose your warranty if you lap your CPU

Also the c2d does run hot but less in the sence because of the overall gain it gives and the difference between load and idle is not all that much.

Also don't worry about the temps too much , its only when you build a system you start noticing that and most of the time the temps are not accurate . You can get a real temp checker to find out.

But honestly don't worry , also remember that your PC very rarely using the core 2 duo will see 100% cpu usage , its only when you run prime and 32 meg pi you see things like that. I have noticed that even when your gaming with the core 2 duo you don't see 100% cpu usage.

You have to remember that it does such high clocks and applications out there are no challenge what so ever. Not even vista cos its design for dual core , the first OS to do that , so you will not see a slowdown with the c2d.
 
killingfield1975 said:
This is my question too about the difference in thermal paste.

because before i used artic silver in my p4 3.0C and was seeing REALLY high temps 46 degrees idle. and i applied and removed and cleaned it properly and i dunno i just had such a hard time getting the low temps. then i got fed up and i used some white generic thermal paste and i got my system to go down to about 31-33 degrees C.
Click to expand...
Wow, so the thermal paste made a 15C difference in your case? Maybe I should try again with another type of paste, I don't think I'm ready to lap the CPU just yet.

And thanks everyone for the info, lots of good stuff here.
 
pdp76 said:
Wow, so the thermal paste made a 15C difference in your case? Maybe I should try again with another type of paste, I don't think I'm ready to lap the CPU just yet.

And thanks everyone for the info, lots of good stuff here.
Click to expand...

yeah i know i was amazed! the instructions told me to us a thin film to the point it was almost transparent. so i did this with artic silver and i had high temps. then i used the generic stuff and i just through a gloob of it on there and achieved much lower temps!!!
 
I experienced lower temps with the default TIM than I did with Artic Silver stuff. I should of just left the default stuff on there. I don't think my processor is totally flat either. I may end up lapping it because I don't really care about warranty to be honest with you.
 
I tried some Arctic Silver, and tried about 3 times reseating the heatsink and apply different amounts, and the most I got my idle temps down to was about 43. I'm giving up on this particular CPU, I'm gonna go in to Fry's and exchange it for another one to see if it's any cooler.

I had another question though, the Arctic Silver instructions said to dab it on the CPU heat spreader, not on the bottom of the HSF. However, when you buy an HSF, sometimes, or most of the times, it already comes with some grease pre-applied to the bottom of the HSF. How come the methods to apply the Arctic Sivler vs the pre-applied grease contradict each other?
 
because the stock heatsinks that intel gives you have no other option except supplying their own tube of thermal paste. I mean technically they could put it on the cpu, but then it just makes it a pain for those that wont be using the stock hsf. Its better this way. But for applying thermal paste to a cpu for use with aftermarket HSF, i see no reason to even consider applying it directly to the HSF and then placing it ontop of the cpu, and hoping none spills during the conversion.
 
nooh said:


on aircooling , dont try this at home :eek:
Click to expand...
On today's episode of Cooking with Nooh we'll be making some delicious beef stroganoff.
First, overclock your CPU and let it warm up to 90C or so by running Orthos.
Then cook finely chopped onions on your CPU for 3 minutes. Add mushrooms; cook just until tender. Remove from CPU.
Brown meat, in batches, in remaining oil. Return meat and vegetables to CPU.
Stir in steak sauce. Heat to a boil; reduce heat to low (shut down Orthos and let idle). Simmer for 15 minutes.
Stir in sour cream. Serve mmediately over noodles.

Anyway, just so I'm not too far OT, I'm getting idle temps of 38 to 42. This is with water cooling and a lapped cpu. Full load temps can be about 55C running orthos large FFTs
and 60C running orthos small FFTs.
 
Nathonius said:
On today's episode of Cooking with Nooh we'll be making some delicious beef stroganoff.
First, overclock your CPU and let it warm up to 90C or so by running Orthos.
Then cook finely chopped onions on your CPU for 3 minutes. Add mushrooms; cook just until tender. Remove from CPU.
Brown meat, in batches, in remaining oil. Return meat and vegetables to CPU.
Stir in steak sauce. Heat to a boil; reduce heat to low (shut down Orthos and let idle). Simmer for 15 minutes.
Stir in sour cream. Serve mmediately over noodles.

Anyway, just so I'm not too far OT, I'm getting idle temps of 38 to 42. This is with water cooling and a lapped cpu. Full load temps can be about 55C running orthos large FFTs
and 60C running orthos small FFTs.
Click to expand...
Wow , i never saw it as cooking the cpu but I thought i would get it ready for Fall and even Winter. I checked to see it could look after itself
:D

I do not thing the temps work , my cpu is lapped and I am using a good TIM also the temps you see there are core temps and My hsf does not get touch hot whilst running these temps.
 
I figure the system would start acting up (crashes, lockups, etc.) if it was too warm.
I don't put too much faith in the value these temperature monitoring programs are giving. I figure you can use them to establish 'normal' idle and load temps specific to your system.
Then, for example, if the temps start creeping up over time then you might check your hardware for things like dust on the HS, malfunctioning fans or pumps, kinks on your water cooling hoses, etc.
 
Nathonius said:
I figure the system would start acting up (crashes, lockups, etc.) if it was too warm.
I don't put too much faith in the value these temperature monitoring programs are giving. I figure you can use them to establish 'normal' idle and load temps specific to your system.
Then, for example, if the temps start creeping up over time then you might check your hardware for things like dust on the HS, malfunctioning fans or pumps, kinks on your water cooling hoses, etc.
Click to expand...
the test of time things at first get better ,TIM settling , ram settling , cpu that is in the first three months , after that you might see build up of dust and what have you.

It worked flawlessly and on the CPU there is a thing called therm trip or something and that will switch itself off , I haven't switched anything like that off so it should be ok.

Also I only tested the waters and thats it , let my TIM settle and I will see better dissipation of heat , I have to say that normally a system if its stable would shut down but this just kept on going and showing all it can. You have to remember Pentium D and what it could do , there is no die difference between them and if that can withstand so much heat so can this.

Intel would like us to believe its different because think if all the people did this and the fact the cpu has a life expectancy of 15 years but on average they only use 4 of them and then it gets too old. I am going to use this CPU for one year or 1.5 max , I want the most of it in the meantime.

I am running at 3.55ghz at the mo with ddr2 960mhz and I am happy with that , I get great benchmarks and great bandwidth , also good latencies. The difference is as much as 30% as from when I was running @3.78ghz with ddr2 840mhz , in bandwidth and the refresh rate or latencies were 130ns when I started at default and its 43ns now, amazing huh.

Will start a thread on that in a few days or so. Also the board is so good that not one time did I have to clear cmos and I have two big fans on the case 250mm which are a god send for my type of system.
 
So I exchanged my processor and got a different HSF. Bad news is, idle temps didn't go down that much, only to about 42C. Good news is, on full load, it only goes to 50C now instead of 65C. Is there something different about Conroes in that mine doesn't have a big difference between idle and full load temps? (only 10C). On my old Northwood, I'd have up to a 25C difference between idle and full load. Any reason for this?
 
What are you using to read the temps? Some temp sensors and monitoring s/w are reading the external CPU temp (which 30C would be fine) and some read the core temp (which 47C would be fine).

Asus Probe @ on the P5W DH reads external CPU temp, for example.

My X6800 idles ~27-30C external temp and ~42-47C core temp.
 
squishy said:
What are you using to read the temps? Some temp sensors and monitoring s/w are reading the external CPU temp (which 30C would be fine) and some read the core temp (which 47C would be fine).

Asus Probe @ on the P5W DH reads external CPU temp, for example.

My X6800 idles ~27-30C external temp and ~42-47C core temp.
Click to expand...
I'm using Intel's Thermal Analysis Tool to read the core temps. But Speedfan, which I'm assuming reads the external temps, are only 1C or 2C different from the core temps.

Anyway, I'm not unhappy with the temperatures anymore, I'm just now wondering why there is only a 10C differential between idle and full load. But maybe that's just how it is and I shouldn't wonder :D
 
That's weird, there's a 10C - 15C difference between the external temp and the core temp on my 6800.

Under load I see the same as you, about a 10C difference.

Edit:

What I see under a slight load

The "CPU" temp matches PC Probe's CPU temp and the core temps match that of CoreTemp.

That being said, I doubt your external and core temps are only 1C or 2C different.
 
the board I have temps are the same as core temps. But that said I wasn't too happy with the temp reader and did think it was fibbing.
 
You must log in or register to reply here.
Back
Top