Abit and Asus CPU Temps: Fact or Fiction?

[Jim]

3-19-04
Written by PF
Edited by Mikki




Abit and Asus CPU Temps: Fact or Fiction?

On every enthusiast forum, when the subject of overclocking is discussed one of the primary limiting factors is CPU temperatures. However, as every oc’er knows, and as is very well documented on the Net, the two favorite manufacturers for the typical enthusiast have a huge discrepancy when reporting CPU temps. Depending on a tweaker’s preference you might hear “Abit read high, Asus is correct”, or exactly the opposite. Here at the BE labs we decided to get to the bottom of this controversy and investigate who’s right, if anyone.


Test Equipment and Configuration


In order to monitor CPU temperatures real time, bypassing the use of software, it was necessary to come up with a way of monitoring those temps at the processor. So after a quick trip to the machine shop, a digital thermal probe and a bit of silicone glue we had our test equipment ready.




http://www.bleedinedge.com/images/pf/One.JPG
First we had a 1/8” hole drilled through the stock Intel HSF assembly 1/16” from the mating surface with the CPU heat spreader.

.

http://www.bleedinedge.com/images/pf/Two.JPG
Once the probe is inserted to mid depth, it would be directly above the CPU heat spreader.



http://www.bleedinedge.com/images/pf/Three.JPG
With the probe in the hole, silicone glue was injected, sealing the probe in the HSF assembly. Once the glue dried the probe essentially was part of the HSF assembly.



http://www.bleedinedge.com/images/pf/Four.JPG
Here’s a picture of the HSF installed in the P4C800-E Deluxe.



http://www.bleedinedge.com/images/pf/Five.JPG
And finally here is a picture of the system running with the digital display mounted in the case. Note upper temp is CPU, lower temp is case.

[Jim]

Now that we have our test equipment laid out, let’s look at the hardware used. Two of the favorite 875 chipset mobos for the enthusiast were used, the Abit IC7 and the Asus P4C800-E Deluxe.


Common equipment for both mobos:
Intel P4 2.6C
Stock HSF modded with thermal probe
Arctic Silver 5 Thermal Paste
2X512 Corsair XMS3200C2
Lian –Li case with 4 Sunon 39CFM fans
BFG FX5900
WD 36GB Raptor on Sata 1
Enermax 431watt PSU
XP PRO-SP1, Detonator 53.03s

And the mobos:
Abit IC7, bios 22. Hardware Doctor and MBM were used to monitor CPU temps through software.

Asus P4C800-E Deluxe, bios 15. Asus Probe and MBM were used to monitor CPU temps through software.


Test Results
Time to put the hardware to work and compare the results reported by Abit’s bundled monitoring software Hardware Doctor and Asus’ bundled software Asus Probe versus the actual temps reported by the digital probe. Mother Board Monitor 5.3.6.0 was used as a control to compare both manufacturer’s software.




We’ll start with both motherboards at stock settings of 200FSB, default VCore at idle.


http://www.bleedinedge.com/images/pf/Abit200IdleDefV.JPG
Abit IC7

http://www.bleedinedge.com/images/pf/Asus2600Idle.JPG
Asus P4C800

Immediately you can see an astounding 17 degree Celsius difference reported by the software on the two motherboards. 45C for the Abit, 28C for the Asus mobo. MBM agreed with both of the manufacturers readings.
Actual temp reported by the digital probe: 32C for the IC7, 33C for the P4C800

Next we loaded both systems running Prime 95 and Super_Pi for 15 minutes.

http://www.bleedinedge.com/images/pf/Abit200LoadDefV.JPG
Abit IC7

http://www.bleedinedge.com/images/pf/Asus2600Load.JPG
Asus P4C800

Under load the IC7 shows 59C, the P4C800 comes in at 40C.
Actual temp reported by the digital probe: 47C for the IC7, 48C for the P4C800

[Jim]

Now we’ll raise the FSB on both systems to 250 and check idle and load temps.


IDLE

http://www.bleedinedge.com/images/pf/Abit250IdleDefV.JPG
Abit IC7

http://www.bleedinedge.com/images/pf/Asus3250Idle.JPG
Asus P4C800

Temps stay consistent with the IC7 at 45C and the P4C800 at 28C.
Actual temp reported by the digital probe: 32C for the IC7, 33C for the P4C800

LOAD

http://www.bleedinedge.com/images/pf/Abit250LoadDefV.JPG
Abit IC7

http://www.bleedinedge.com/images/pf/Asus3250Load.JPG
Asus P4C800

Under load the IC7 shows 61C, the P4C800 comes in at 45C.

Actual temp reported by the digital probe: 51C for the IC7, 53C for the P4C800

[Jim]

Finally we left the FSB at 250 and changed the VCore from default to 1.6V and checked idle and load temps.

IDLE

http://www.bleedinedge.com/images/pf/Abit250Idle1.6V.JPG
IC7

http://www.bleedinedge.com/images/pf/Asus3250VC1.6Idle.JPG
P4C800

At idle, 250FSB and 1.6Vcore the IC7 reported 48C, the P4C800 came in at 31C.
Actual temp reported by the digital probe: 35C for the IC7, 36C for the P4C800


LOAD

http://www.bleedinedge.com/images/pf/Abit250Laod1.6V.JPG
IC7

http://www.bleedinedge.com/images/pf/Asus3250VC1.6Load.JPG
P4C800

Under load, 250FSB, 1.6VCore, the IC7 reported 66C, the P4C800 came in at 48C.
Actual temp reported by the digital probe: 53C for the IC7, 56C for the P4C800

[Jim]

Let’s put it all together


Idle, 200FSB, Default VCore

Abit: 45C
Asus: 28C
Actual: 32C/33C

Load, 200FSB, Default Vcore

Abit: 59C
Asus: 40C
Actual: 47C/48C

Idle, 250FSB, Default Vcore

Abit: 45C
Asus: 28C
Actual: 32C/33C

Load, 250FSB, Default Vcore

Abit: 61C
Asus: 45C
Actual: 51C/53C

Idle, 250FSB, 1.6Vcore

Abit: 48C
Asus: 31C
Actual: 35C/36C

Load, 250FSB, 1.6Vcore

Abit: 66C
Asus: 48C
Actual: 53C/56C


Conclusion

The numbers definitely speak for themselves. The Abit software reported temps anywhere from 10-13 degrees Centigrade high, the Asus software read anywhere from 5-8 degrees Centigrade low as compared to the digital probe. There are definitely more efficient cooling solutions available to the enthusiast on the market today, but the readings should remain consistent whatever the hardware used for cooling. Indeed, we only performed a mild overclock with a modest VCore boost using stock cooling, but what we were searching for here is the discrepancy in CPU temp reporting by the 2 companies. It would seem a simple fix for the software to be updated by both manufacturers to more consistently reports CPU temps. Actual temps were virtually identical and the small difference can probably be attributed to the Asus mobos tendency to over volt - of course that’s a whole other discussion.
The BE hopes you find these numbers enlightening and will continue to tweak your hardware to it’s maximum potential while also continuing to use caution when stressing your hardware.

[Flamer]

PF,

This is fantastic info :thumb: It is kinda funny to see how far off the sensors are. Many have asumed for a while that asus reads low and abit high,this proves it:) I would love to see this done to a few more boards :thumbs: Great work bud.

[ReDoCeD]

Good stuff proff :beer: i have my sensor under the procceser so would this be wrong? i get about 10 to 12 c less then abit reports idle 26c load 41c

[Mikki]

This subject has certainly come up a few times over the years.... :p :rolleyes: Fantastic testing PF, thanks for finally giving us some facts to go with our assumptions... :wpleased:

That modded heatsink is a very handy tool, I may have to do that... :chit: :bow:

[Moonwizard]

Awesome. Now I can change the mbm5 thingie and lower the temps by 10 degrees. :D Nice testing. :D

[dbwillis]

Great work PF!!

Its nice to know that everyones guesses on the differences were somewhat right and now we can be sure!

[RotorHead]

Great Work PF :thumbs:

This kind of information is a must have for anybody using those boards and wanting tweak there system.

[Paragon]

Jim :thumbs:

NOW THAT'S WHAT I'M TALKING ABOUT...

This is probably one of the most useful reviews I've ever read, it pinpoints exactly the issue being examined and gets right to the root of the problem without a lot of redundant garbage that a lot of reviewers include to cloud the matter they are investigating.

Jim I would give this review an 11 on a 10 scale...Thanks so very much for a great, pertinent review that will be extremely useful to all that read it! :rock: :beer: :rock:

[Hawk]

Awsome work PF. . . .

[Jim]

Thanks for all your kind words. :o

samcolt , yeah I'd like to ck some other mobos too. Who's going to send them to me? :lol:

ReDoCeD , sounds like your setup is working good for you as-is. :thumbs:

[Jamesbond2k]

Nice work Prof. This confirms the results of my own test :thumbs:

[ThugsRook]

:beer: :beer: :beer:
:beer: :beer: :beer:

(six pack)

[Hawk]

:beer: :beer: :beer:
:beer: :beer: :beer:

(six pack)



:lol: :lol: :lol: awsome. . . .

[muzz]

Excellent job, that probe was mighty close, so it should be VERY close to actual temps.
Send that to Abit and Asus and let them sort it out!!

:beer:

[ccokeman]

Its a great test just 1 thing bothered me. If MBM and asus probe are run at the same time the data gets corrupted.Im not sure if that is the case with abits software but it has happened to me with my asus boards. Plz dont take this the wrong way just offering my thoughts . :bonk:

[Jim]

Its a great test just 1 thing bothered me. If MBM and asus probe are run at the same time the data gets corrupted.Im not sure if that is the case with abits software but it has happened to me with my asus boards. Plz dont take this the wrong way just offering my thoughts . :bonk:

ccokeman , welcome to BE. :welcome:
I was aware of this issue and did verify temps individually, but also wanted to show all the data in one screenie. :wave:

[TheDude]

Great review, but I couldn't find one vital piece of information.....which thermometer and thermal probe was used and was it recently calibrated?
I am switching from the Asus deluxe to the abit Max3 today so this info was right on time. Thanks!

[Gaspar]

Hey Dale Welcome to the Forum I think you know who I am

[Ehryk]

This is probably the best temp test I have seen yet to date, great job :rock:

With the new 1.6 BIOS for the IC7-MAX3 that claims to have calabrated the temps... whats the chances of getting you to do one more round and show us where its at now? :o I know many of us would deeply appreciate it. :chit:

E

[Jim]

Ehryk , welcome to the BE. :welcome:
And thanks very much for the interesting update on the latest IC7 series bios. I don't currently have the IC7 on my desk but will make an effort to test one with the new calibrated bios. :beer:

[Ehryk]

Ehryk , welcome to the BE. :welcome:
And thanks very much for the interesting update on the latest IC7 series bios. I don't currently have the IC7 on my desk but will make an effort to test one with the new calibrated bios. :beer:

Thanks :thumbs: :beer:

E

[falltime]

what is the next best place to put the temp probe if you dont want to drill a hole through your heatsink? I have mine taped to the bottom of the heatsink directly adjacent to the die. is that sufficient?

[pointreyes]

This is probably the best temp test I have seen yet to date, great job :rock:

With the new 1.6 BIOS for the IC7-MAX3 that claims to have calabrated the temps... whats the chances of getting you to do one more round and show us where its at now? :o I know many of us would deeply appreciate it. :chit:

E

The calabrated temps only work for the Prescott procs. Got that info from other excited and then downtrodden individuals. Note that my Asus board seems to provide accurate readings compared to my Koolance waterblock temp probe. I wonder if that is due to using a Prescott on a 'Prescott' based bios as well?