This is somewhat related to this question but at a less serious level - one that causes frustration rather than permanent damage.
From programs like CoreTemp and Speedfan, it is clear that there are ways to read the critical temperature and voltage information from a system. Is there any means of (maliciously) writing parameters into the same monitoring systems so that the BIOS (or whatever is responsible) could be fooled into shutting a system down believing it was dangerously over-voltage or over-temperature? If so, are there any viruses known to operate this way?
I ask because I have a system (ASUS A6U laptop) which was having problems that appeared to be either a virus or a thermal shutdown issue. The essential problem was a strong tendency to fail to boot into Windows XP, which usually occurred immediately after an error reported by lsass.exe (a frequent target of viruses). But moments later the system would always suddenly shut down in a way suggestive of thermal issues. Still, the system would run fine in BIOS for long periods of time, or go through the Windows setup process to a reasonably predictable point, and both suggested the issue had to be dependent on software to some degree.
Interestingly, after restoring the laptop to factory settings on the boot partition (but not reformatting a data partition), it worked fine for a couple of days, lending further weight to the virus/software issue hypothesis. But then it reverted to the previous behaviour. And in the lead up to having boot problems again, it had started to report very strange temperatures through the ASUS NB Probe software.
So I'd like to understand how voltage and temperature are (typically) monitored (are sensors internal or external to CPU; does it happen in CPU, BIOS or operating system; are there read-write parameters; etc) so that I can rule in or out the possibility of virus activity.