CPU/MB temps

[Eric]

Is it normal for the CPU to run cooler than the motherboard? I have an
Athlon 64 3400+ and an ASUS K8N board and the probe says the cpu generally
runs about 30-33 degrees C, but the MB is usually a few degrees warmer than
that. I know those are low temps and it probably doesn't matter, but I'm
just curious if it's normal or not. Thanks!

 

[kony]

Is it normal for the CPU to run cooler than the motherboard? I have an
Athlon 64 3400+ and an ASUS K8N board and the probe says the cpu generally
runs about 30-33 degrees C, but the MB is usually a few degrees warmer than
that. I know those are low temps and it probably doesn't matter, but I'm
just curious if it's normal or not. Thanks!

When CPU is in a power managed state it is certainly
possible, even more common with today's southbridges which
run hotter than those of a few years ago. The temp sensor
for the system is typically in the north or southbridge, so
it's not actually a "system" ambient temp... would be
impossible for the CPU to actually be cooler than the system
ambient air temp but certainly possible for it to be cooler
than another chip.

So, normal compared to similar technology, abnormal compared
to older Athlons, somewhat common on even older Coppermine
based platforms. It's nothing to be alarmed about providing
you believe the sensor is accurate or at least that all
temps reported are below a certain threshold, say 70C (just
in case there were some bug causing sensor output to be
reported as, assigned to the wrong device), and no
instabilities that seem temp-related.

 

[Mike Walsh]

The CPU normally runs warmer than the motherboard. The temperatures might be reported incorrectly. My Soyo motherboard with a Phoenix BIOS incorrectly reports CPU temperature as motherboard temperature and motherboard temperature as CPU temperature.

 

[Eric]

The CPU normally runs warmer than the motherboard. The temperatures might
be reported >incorrectly. My Soyo motherboard with a Phoenix BIOS
incorrectly reports CPU temperature as >motherboard temperature and
motherboard temperature as CPU temperature.

how did you figure out it was switching the readings?

 

[kony]

how did you figure out it was switching the readings?

There's a relatively easy way to tell, unplug the CPU
heatsink fan while watching the temp readings. It'll stay
cool enough to be safe, long enough to note which is rising,
then power off the system and plug the fan back in... just
don't run it for more than a few dozen seconds without the
fan.

 

[Mac Cool]

kony:

There's a relatively easy way to tell, unplug the CPU
heatsink fan while watching the temp readings.

There's an easier way: run something CPU intensive such as Prime95 torture
test. The CPU temp will rocket up while the motherboard temp should remain
relatively stable.

 

[Mike Walsh]

Without unplugging any fans; the CPU temperature will go up during high CPU usage and the motherboard temperature will go up when running with the case panels off because air is not directed over the motherboard.

There's a relatively easy way to tell, unplug the CPU
heatsink fan while watching the temp readings. It'll stay
cool enough to be safe, long enough to note which is rising,
then power off the system and plug the fan back in... just
don't run it for more than a few dozen seconds without the
fan.

--

When replying by Email include NewSGrouP (case sensitive) in Subject

Mike Walsh
West Palm Beach, Florida, U.S.A.

 

[kony]

kony:


There's an easier way: run something CPU intensive such as Prime95 torture
test. The CPU temp will rocket up while the motherboard temp should remain
relatively stable.

I agree that can work but other times it may not. For
example I've an XP Barton with an SLK-900 plus a fan on a
thermal sensor. Full load and idle temps are within 2-3
degrees of each other. Then again, it's not 30C either,
stays closer to 38C to minimize noise. The lower the idle
temp the more likely there'd be noticable difference.

 

[Mac Cool]

kony:

I agree that can work but other times it may not. For
example I've an XP Barton with an SLK-900 plus a fan on a
thermal sensor. Full load and idle temps are within 2-3
degrees of each other. Then again, it's not 30C either,
stays closer to 38C to minimize noise. The lower the idle
temp the more likely there'd be noticable difference.

Very impressive, but unless the OP has a significant HSF like yours, I
imagine his CPU temps will increase enough to be obvious.

 

[Mac Cool]

kony:

I've an XP Barton with an SLK-900 plus a fan on a
thermal sensor. Full load and idle temps are within 2-3
degrees of each other.

I'll be darned... I've never run Prime95 on my current machine (P4
Northwood) so I tried it after reading your post. My CPU temp only went up
2 degrees on the maximum heat test, from 33 to 35 with factory heatsink.
Big improvement since the last time I ran Prime95 torture test which used
to double the temps on my old Thunderbird.