What would you trust?

[KevCritch]

I have a xp3200 running on a A7N8X, 3 case fans and a AKASA 827 Heatsink/Fan.

My bios hardware monitor tells me cpu temp 50-57c and M/B temp 43c

But the latest version of Asus PCProbe is telling me my cpu is 30c and m/b is 43c

What would you trust and why? answers on a postcard!

 

[Zepper]

Are they both reading from the same source? I've heard that some mobos can read from both an external CPU sensor and the internal diode.
.bh.

:moon:

 

[yadda]

OK I couldn't think that the CPU could be 7-14 degrees warmer. Notpossible.

Y

 

[KevCritch]

Is it possible for the cpu to run at a lower temp than the m/b temp with air cooling?

 

[Zepper]

Very unlikely.
.bh.

 

[lobadobadingdong]

what's speed fans temp detectors say?

 

[KevCritch]

if you mean the cpu on the heatsink its at 2300rpm,

To be honest the system works fine no shutdowns, in the future i might invest in a new heatsink and fan but for now 50-60 is high but acceptable for me.

Thanks for your help

Kev

 

[Fern]

Originally posted by: KevCritch
I have a xp3200 running on a A7N8X, 3 case fans and a AKASA 827 Heatsink/Fan.

My bios hardware monitor tells me cpu temp 50-57c and M/B temp 43c

But the latest version of Asus PCProbe is telling me my cpu is 30c and m/b is 43c This looks wrong. I've never seen a cpu cooler than the mobo. Looks like the sensors are "switched" under the probe utility

What would you trust and why? answers on a postcard!

Fern

 

[BonzaiDuck]

This forum has visited this issue before. The explanation has to do with hyper-threading under a running OS, and the frequency of temperature sampling under the BIOS monitor.

I may not have this explanation perfectly correct, but under the BIOS monitor in the system's pre-boot state, the only thing the processor can do in that state is "sample, sample, sample" from the thermal monitor. I think -- I THINK the explanation involves the sampling process itself -- and the lack of hyper-threading in that state -- actually causing the CPU to heat up a bit.

The temperatures reported under the OS, with hyperthreading operative, are as reliable as the motherboard software like "ASUS Probe" or "Intel Active Monitor" or for that matter MBM5. You will always -- ALWAYS -- have BIOS monitor temperatures which are higher than temperatures with the OS and hyperthreading.

Try attaching a thermal sensor to the CPU heatsink as near as possible to the processor without squishing it between the heat-sink and processor cap, then connect it to a front-panel backlit temperature display. Some heatsinks will be a few degrees hotter than the processor temp reported by MBM5-etc., and some heatsinks will be a few degrees lower. The only explanation for that would be the mass and thermal properties of the heatsink base, and the airflow around the base. But the discrepancy -- note that I say it seems to vary with heatsink while using the same thermal monitor -- has nothing to do with the discrepancy reported in BIOS.

 

[jkresh]

asus probe is more likely to be correct then bios, though it may have switched motherboard and cpu temps, you could try running prime95 and seing whcih temp goes up more (the one that increases more is likely to be the cpu) if niether increases (and the temps you are posting are from idle) then something is wrong with the sensors

 

[BonzaiDuck]

asus probe is more likely to be correct then bios, though it may have switched motherboard and cpu temps

I've never seen that happen. Not if the chipset and SMBus drivers were correctly and cleanly installed immediately after the OS setup.

Using the same ASUS P4P800 mobo I have now, I had a 2.4C Northwood running in it for 4-and-a-half months, and the CPU temperature actually stayed 2F below the motherboard temp as reported. But you knew which was which when you put the system under load -- the CPU was being correctly monitored by the Probe software . . .

 

[HardWarrior]

Originally posted by: Zepper
Are they both reading from the same source? I've heard that some mobos can read from both an external CPU sensor and the internal diode.
.bh.

:moon:

On the A7N8X-dlx there's the diode sensor (in the CPU) and a bulb sensor in the ZIF (socket) under the CPU. The difference bettwen the two is like 3c on my box with the socket being warmer than the diode.

 

[Fern]

On the A7N8X-dlx there's the diode sensor (in the CPU) and a bulb sensor in the ZIF (socket) under the CPU. The difference bettwen the two is like 3c on my box with the socket being warmer than the diode.

That sounds odd... AFAIK, the diode should report higher temps than a socket?

Fern

 

[CheesePoofs]

Originally posted by: HardWarrior
On the A7N8X-dlx there's the diode sensor (in the CPU) and a bulb sensor in the ZIF (socket) under the CPU. The difference bettwen the two is like 3c on my box with the socket being warmer than the diode.

This is not possible, the die is the part of the cpu where all of the heat is actually created, so it will have to be than the sensor in the socket, because the heat has to go through a few layers of PCB and plastic before it can reach the socket's sensor.

 

[HardWarrior]

Originally posted by: Fern

On the A7N8X-dlx there's the diode sensor (in the CPU) and a bulb sensor in the ZIF (socket) under the CPU. The difference bettwen the two is like 3c on my box with the socket being warmer than the diode.

That sounds odd... AFAIK, the diode should report higher temps than a socket?

Fern

It used to, before I moved to water. I can't explain it either.

 

[Fern]

That's interesting. I can't think of a reason for that. But I've never played with water cooling so don't know much about it.

 

[HardWarrior]

Originally posted by: Fern
That's interesting. I can't think of a reason for that. But I've never played with water cooling so don't know much about it.

I was suprised as well. I've noticed also that my computer room is a bit warmer that it was before I went to water. Someone at procooling told me that this is because thermal energy that had been locked up in my case is now being exhausted.