Is Motherboard Monitor 5 more accurate then the monitor in your mobo's bios?

[Compddd]

Mother Board Monitor 5 says my CPU is 34C idle, when I goto my BIOS, it says CPU is 38C Why the difference in temp, and which one is the correct one?

 

[Megatomic]

Originally posted by: Compddd
Mother Board Monitor 5 says my CPU is 34C idle, when I goto my BIOS, it says CPU is 38C Why the difference in temp, and which one is the correct one?

I don't think MBM5 can be more accurate than the BIOS monitor as it gets it's signals from the BIOS. But what might be happening is that the CPU cools down a bit when you exit windows and reboot into the BIOS. But that's just my speculation.

 

[Texun]

I don't have the answer but I can give it a shot and maybe somebody can finish it later. What cpu are you using and are you running WinXP?

IIRC Windows XP has a typs of software cooling that is used when the cpu is not under load. This might explain why Win temps are lower than the BIOS. Can somebody help me out with the rest? Ever hear of Vcool? It does the same thing. I used to run Vcool and had great results but it does nothing for me now. Maybe it's WinXP or the chipset. It's cool enough so I no longer worry about it.

 

[Compddd]

My system board temp is 32-33C idle, I need some good intake/outake fans I think, I read somewhere that your system board temp shouldn't be over 29-31 idle

 

[ugh]

If I'm not mistaken, when you're in the BIOS, the CPU is not really idle coz there isn't any HALT instruction sent to the CPU. This means the CPU will be hotter as if it's processing stuff. When you've booted into the OS, the OS will take care of the HALT instructions if there aren't any stuff to process.

 

[Zepper]

And mobo monitor gets its data from the health chip, not the BIOS. So that means that even though they are reading the same data, one can display the data better than the other depending on the chosen algorithms.
. As another said, readings can also be diff due to diff conditions when using the CMOS setup program than when in the OS. They really SHOULD be of the same accuracy.
.bh.
:beer: time!

 

[seismik]

My system board temp is 32-33C idle, I need some good intake/outake fans I think, I read somewhere that your system board temp shouldn't be over 29-31 idle

I'd be curious to know where you read that. My case temp is up around 36-37 idle, and my CPU is around 38 idle (from Soltek motherboard monitor which I suspect measures a touch high). I'm not worrying about it much, case temp doesn't get above 40 even when the CPU is stressed and I'd be pretty surprised if that kind of temperature has any negative effect on the life or performance of (non-overclocked) components. You're running some pretty hot (literally) components, like the 9800 Pro and a P4C that will naturally raise your case temperatures alot.

If someone disagrees with this, please let me know and help save the life of my computer!

 

[Macro2]

They are both correct. Assuming they are both reading the same thermal diode.

When you are in the BIOS the CPU is running full on. When in Windows it's being idled. The BIOS temp will real higher.

With the temps you mentioned I'd be a happy camper.

 

[Megatomic]

Originally posted by: Zepper
And mobo monitor gets its data from the health chip, not the BIOS. So that means that even though they are reading the same data, one can display the data better than the other depending on the chosen algorithms.
. As another said, readings can also be diff due to diff conditions when using the CMOS setup program than when in the OS. They really SHOULD be of the same accuracy.
.bh.
:beer: time!

So MBM is able to access the Winbond chip directly without going through BIOS? That just doesn't seem right to me. How is it able to do this?

 

[Zepper]

MBM has customizations for each health chip. The new version that was posted on the 9th (see my post in this section) has support for the new NVidia health chip.
Any good diagnostic program can access the various IC's on the mobo directly. In most cases the BIOS isn't used at all once Win has loaded.
I haven't read the specs on health chips yet, but it's a good bet that they put their data into an I/O port or ports at a given address(es). All the MBM has to know how to do is find the right I/O port address(es) and read/translate the bit string(s) found there.
Pretty simple.
.bh.
Where's the :sun: ?

 

[KF]

>So MBM is able to access the Winbond chip directly without going through BIOS?
> That just doesn't seem right to me. How is it able to do this?

This is actually a complicated question. AFAIK (and I am reasonably sure) there are no BIOS system calls that a program can make to get a report this type of info.

Regardless, operating in protected mode as Windows does, an application program is not allowed to directly access the hardware. The OS intercepts such instructions and from there may do whatever it pleases. The OS could do nothing. It could issue an error warning. It could decide to actually run a progam at the correct privlege level that does access the hardware. Such an OS program would be "wrapped" with appropriate protection measures. As I understand it, there is a program module (like a driver) included with MBM that does what is necessary for the OS to allow access to IO ports. When things are done this way, a program supposedly can never lock up the computer. Different temperature/health chips operate differently, and MBM has INI files with info about how to operate many of them.

On the mobos have I have owned or had access to, the BIOS temps are high compared to the Windows idle temperature. Modern X86 chips (I have been told) normally power down parts that are not in use, but it appears that this is partially disabled while you are settiing things in the BIOS.

To be specific about accuracy: you are going through several levels of translation. There is a diode or thermistor which is set up to produce a voltage proportional to the temperature. Depending on how this is done, and how well, you get a more or less accurate analog to the temperature. With measuring instruments there would be some calibration method to get them close, but on mobos I am sure there is not. Obviously thermometers only measure the temperature at their exact location, so some offset or compensation may be arithmetically applied by MBM or the BIOS to (supposedly) derive the CPU temperature. The specialty chips that convert the analog voltage to a digital reading may not have particularly accurate A to D converters, so a one digit change is not necessarily one degree C change.

At the end, any digital system can't be more accurate than one step, which in this case is one degree C. But this is the least of the uncertainties involved.

To sum up, there does not seem to be any reason to think mobo monitoring hardware is very accurate. And mobo makers make no claim about the level of accuracy.