CPU idle vs. burn... no temp difference. Normal?

Cooling Processors quietly

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Centaur
Posts: 19
Joined: Fri Jul 29, 2005 3:15 pm

CPU idle vs. burn... no temp difference. Normal?

Post by Centaur » Sat Aug 20, 2005 9:45 am

I’ve just finished building my new quiet PC and have noticed the CPU temp barely changes from the idle temp when it is stressed. At idle, Speedfan shows 46c, when I stress with CPUBurn (set to K7) the highest it goes after 1/2 hour is 48. I’ve checked the heatsink to the touch after this and it’s not even close to hot, more like lukewarm. So I’m wondering if perhaps CPUBurn isn’t doing it’s thing (it shows the CPU at 100% utilization while running), or if the temps aren’t being reported correctly which I think the touch test disproves. Maybe I just got lucky. :)

Regardless, I’m curious if anyone else has had this happen with their setups and any thoughts/comments. After a lot of reading seems like everyone's temps go up significantly under stress. Applicable system specs:

ASUS A8N-SLI Premium
Antec SLK3000B case
3200+ Venice
Scythe Ninja SCNJ-1000 CPU cooler
Artic Silver 5 applied (a bit more than recommended, long story)
ATI Radeon x800XL 256meg video w/VF700ALCU cooler
Seasonic S12-430 PSU
Corsair Valueselect DDR ram – 2 gig (512x2x2)
Yate Loon 120mm orange fans at rear (exhaust) and front (intake) running at 900rpm
Running fanless on CPU cooler!
Not running Cool N Quiet
No tweaks/overclocking done, running stock on all components
Room temp is 27c

Overall I’m pleased with the temps. Would like to see the idle temp a little lower but with the stress temp not changing guess it doesn’t matter. FWIW, the video card runs at 39 constant, cooler fan set to 7 volts. Haven’t tried stressing it yet. Motherboard temp stays at 40.
Last edited by Centaur on Sat Aug 20, 2005 9:49 am, edited 1 time in total.

qviri
Posts: 2465
Joined: Tue May 24, 2005 8:22 pm
Location: Berlin
Contact:

Post by qviri » Sat Aug 20, 2005 9:48 am

Aren't Athlon 64s K8, rather than K7?

Centaur
Posts: 19
Joined: Fri Jul 29, 2005 3:15 pm

Post by Centaur » Sat Aug 20, 2005 9:52 am

qviri wrote:Aren't Athlon 64s K8, rather than K7?
Correct, but CPUBurn does not have a setting for K8. I read elsewhere that the K7 setting would be sufficient for testing a K8. Perhaps not? Is there another good/recommended CPU stress test I can run?

Qwertyiopisme
Posts: 237
Joined: Tue Nov 11, 2003 6:48 am
Location: Gothenburg, Sweden
Contact:

Post by Qwertyiopisme » Sat Aug 20, 2005 9:56 am

Prime95 perhaps?

Mystr_Byrnz
Posts: 104
Joined: Fri Jul 08, 2005 12:02 am

Post by Mystr_Byrnz » Sat Aug 20, 2005 2:13 pm

Why aren't you running cool n quiet?

I have the same CPU with a Zalman 7000AlCu and my idle temp is MUCH lower than yours. around 33c. Sure it's not silent, but it is by no means loud.

StarfishChris
Posts: 968
Joined: Fri Jan 07, 2005 7:13 pm
Location: Bristol, UK
Contact:

Post by StarfishChris » Sat Aug 20, 2005 2:58 pm

I triple-boot and my Windows 98 installation has no loading programs nor a reasonable power saving function in the OS (AFAIK), but the temperature drops from 44° to 37° with C&Q off. It's a 3200+ Venice like yours, running at 1.15v. I can't think why there is such a small difference...

vertigo
Posts: 647
Joined: Sun Apr 17, 2005 6:09 am
Location: UK

Post by vertigo » Sat Aug 20, 2005 3:00 pm

There's something seriously wrong there. At the moment, my CPU only raises 2C from idle to load, but that is because it is a 3000+ running at half speed, and on only 0.95v. The XP/90 is overkill for the heat it produces (17W or so).

However, you have stock settings and a fanless heatsink. Something is seriously not right...

Centaur
Posts: 19
Joined: Fri Jul 29, 2005 3:15 pm

Post by Centaur » Sat Aug 20, 2005 3:58 pm

Okay, I enabled Cool N Quiet per the ASUS manual... setup in BIOS and set power management feature in XP accordingly, but had no affect at all. With the CNQ monitor software running my CPU frequency never changes from 2000 mhz nor does the voltage change from 1.3750. Apparently something is not setup correctly or I'm doing something wrong.

I have 3 hard drives in my system, all SATA's. Speedfan typically shows their temps around 44 on average. With the intake fan blowing across those, and then the exhaust taking it out across the CPU perhaps that is contributing to my high idle temps but am still confused why no real change under stress. I do live in Arizona and it's quite warm today... my last Dell always struggled during summer so that could be part of it as well.

Finally, I installed Prime95 and ran... it had no affect at all on the temps whereas CPUBurn was able to kick it up a few.

vertigo
Posts: 647
Joined: Sun Apr 17, 2005 6:09 am
Location: UK

Post by vertigo » Sat Aug 20, 2005 4:01 pm

What windows version are you running?

If you are in 9x/ME, go there, get rain10.zip, run that proggy and see if it makes a difference.

Centaur
Posts: 19
Joined: Fri Jul 29, 2005 3:15 pm

Post by Centaur » Sat Aug 20, 2005 4:10 pm

vertigo wrote:What windows version are you running?
I'm on XP, SP1 so guess I don't need that. Thanks for the link though, and for the replies.

One thing I thought of: When I applied the Arctic Silver 5 I may have overdone it. The Ninja is kinda hard to install, after the first application and seating it down I accidentally lifted it off the CPU "breaking" the seal. Without wiping off what was already there I applied another BB sized blob of the stuff and re-seated. Not sure if that qualifies as overdoing it but it's something I've wondered about as I've heard too much is not a good thing.

Mystr_Byrnz
Posts: 104
Joined: Fri Jul 08, 2005 12:02 am

Post by Mystr_Byrnz » Sat Aug 20, 2005 4:59 pm

AFAIK that should not have caused any problem(the AS5 thing you did)

As long as nothing got between your CPU and the HSF when you lifted it you should still get a good seal from what I've been told.

Oliver
Posts: 116
Joined: Sun Feb 13, 2005 8:01 pm

Post by Oliver » Sat Aug 20, 2005 7:44 pm

Centaur:

When you break the seal, you can get bubbles. But if you lifted it off, then reapplied a dab in the middle, without wiping down the surface first, then just put the heatsink back --- well that is an interesting question as to whether you would get a bubble. My inexperienced guess would be you probably would not get a bubble in the center portion where it counts, but probably would get some around the perimeter. But I really have very little experience , so that is a guess. You migh get some of it leaking around the edges as you now have 2 dabs worth on it :)

As far a cool-n-quiet goes. Well not only do you have to set it in the BIOS, but then if you are using something like Windows xp, you have to go to the advanced power managment panel that is part of the operating system, and then you have to check some low power configuration box before the Cool-n-quiet kicks on. When it does kick in , then you will be able to observe it with watever tools your mobo manufacturer gives you to do so. I forget the specific box you have to check under windows and the specific name of the panel, but that is what you neglected to do. Do that then it will kick in.

EDIT:

I TAKE BACK MY FIRST PARAGRAPH. I THINK YOU GOT BUBBLES IN THE CENTER TOO! REAPPLY ACCORDING TO DIRECTIONS FROM MANUFACTURER.
Last edited by Oliver on Sat Aug 20, 2005 8:51 pm, edited 1 time in total.

Oliver
Posts: 116
Joined: Sun Feb 13, 2005 8:01 pm

Post by Oliver » Sat Aug 20, 2005 7:50 pm

By the way , a tiny bubble above the cpu die itself , is not the best thing, because then that tiny portion of the cpu gets hotter than the surrounding portions of the cpu. So one portion of the cpu may be getting much hotter than where the thermal diode is embedded in the cpu. But probably since you are using a cpu that is capped with that heatspreader (mobile cpu's don't have that feature), that thin piece of metal itself evens out the heat of the cpu right behind it (a little bit). So desktop cpu is probably less sensitive to bubbles than mobile cpu would be. Just guessing though as I have little experience.

darthan
Posts: 237
Joined: Sat Apr 30, 2005 1:28 pm
Location: San Francisco

Post by darthan » Sat Aug 20, 2005 8:02 pm

I have a 3200+ on an A8N-SLI with a Ninja as well and it idles at 30C (in a 20C room) and goes up to 42C when running CPUBurn. This is according to the AsusProbe monitoring utility that came with my board. I suggest checking using this as I have definitely seen speedfan fail to report accurate temperatures before.

It is also possible your Winbond temperature sensor onboard might be broken or misreporting. I'd suggest contacting Asus tech support if the AsusProbe doesn't show something more believable.

alleycat
Posts: 740
Joined: Sun Oct 20, 2002 10:32 am
Location: Melbourne, Australia

Post by alleycat » Sat Aug 20, 2005 8:37 pm

There is, without doubt, a problem with the thermal interface between the CPU and the heatsink. Either the HS isn't mounted properly, or there is something wrong with the way the thermal grease has been applied. I suggest you remount the HS. I have heard a few reports that AS5 is difficult to work with, although I haven't tried it myself. I recently built an A64 system using a Ninja HS. I found the thermal grease supplied with the Ninja to be very easy to use.

Oliver
Posts: 116
Joined: Sun Feb 13, 2005 8:01 pm

Post by Oliver » Sat Aug 20, 2005 8:49 pm

i take back what I said above.
I believe you have bubbles above the cpu die , in the center in addition to the perimeter.
You see, if you have a bubble there to start (which you certainly introduced when you broke the seal), then those bubbles in the center will undoubtably get mixed in when you put your new dab in the center.

You should start from scratch.
Reapply according to manufacturer's directions.

Centaur
Posts: 19
Joined: Fri Jul 29, 2005 3:15 pm

Post by Centaur » Sun Aug 21, 2005 7:34 am

alleycat wrote:There is, without doubt, a problem with the thermal interface between the CPU and the heatsink. Either the HS isn't mounted properly, or there is something wrong with the way the thermal grease has been applied. I suggest you remount the HS.
I was beginning to think that's what I was going to have to do (unfortunately) though am curious about one thing. If the temp of the components around the CPU are in the 42-44 range, a CPU temp of 48 seems to make sense, the thought being the surrounding components are driving up the temp of the CPU. I have a feeling a remount might bring it down a couple but am wondering if it's worth the effort. I've been reading SPCR for only a few weeks, not sure if a CPU temp is typically the same or less than the components, or if the components can drive up the CPU temp just by being in the same airflow. The Ninja is an effort to install... had a hard time doing it outside the case. Not looking forward to doing it inside the case! :(

Thanks everyone for the replies and info!

alleycat
Posts: 740
Joined: Sun Oct 20, 2002 10:32 am
Location: Melbourne, Australia

Post by alleycat » Sun Aug 21, 2005 11:36 pm

The temps of the other components are largely irrelevant. For example, I have a low-airflow P4 system with a CPU idle temp in the low 40s, but the northbridge, which is immediately below the CPU HS, consistently reports 60degC and is very hot to the touch.

It doesn't seem that your temperatures are at dangerous levels, so maybe you don't really need to correct the problem. I'll admit that mounting the Ninja inside the case could be a pain, however, it is not a difficult HS to install. I've written about how I did it in another thread. The hardest part would be cleaning the AS5 off everything.

ushac
Posts: 19
Joined: Mon Jul 25, 2005 4:29 am
Location: Stockholm, Sweden

Post by ushac » Mon Aug 22, 2005 6:18 am

About cool n quiet, did you also install the amd cpu-driver? I got an asus CnQ monitoring utility with my motherboard which I can use to check how the clock frequency goes up and down in real time. I enabled CnQ in BIOS and set power management to minimum in windows xp, but I didn't change anything until I found that driver...

After reading this thread I have realized I should fix my heat sink installation too. It's just such a hassle to get it out...

/ushac

Centaur
Posts: 19
Joined: Fri Jul 29, 2005 3:15 pm

Post by Centaur » Mon Aug 22, 2005 6:30 am

ushac wrote:About cool n quiet, did you also install the amd cpu-driver? I got an asus CnQ monitoring utility with my motherboard which I can use to check how the clock frequency goes up and down in real time. I enabled CnQ in BIOS and set power management to minimum in windows xp, but I didn't change anything until I found that driver...
/ushac
Thanks for the tip, I think that may be the problem because I don't remember loading that driver. Will give it a shot.

Oliver
Posts: 116
Joined: Sun Feb 13, 2005 8:01 pm

Post by Oliver » Mon Aug 22, 2005 11:04 am

darthan wrote:I have a 3200+ on an A8N-SLI with a Ninja as well and it idles at 30C (in a 20C room) and goes up to 42C when running CPUBurn. This is according to the AsusProbe monitoring utility that came with my board. I suggest checking using this as I have definitely seen speedfan fail to report accurate temperatures before.

.
Are you running your Ninja fanless, or do you have a slow speed fan on it?

Color Theory
Posts: 17
Joined: Tue Aug 16, 2005 1:16 pm

Post by Color Theory » Mon Aug 22, 2005 12:59 pm

I can confirm that CnQ won't work until you download the "AMD Athlon Processor Driver" from AMD. Here's the page for the dual cores. There's a link to the single core page on the right.

http://www.amd.com/us-en/Processors/Pro ... 18,00.html

Post Reply