Well, that at least confirms my suspicions that there's something seriously amiss here.

I'll open a support call... not that I expect them to do anything useful, based on experience to date.

Specfically which tools are you using to measure?
The system comes with some nVidia monitoring software installed. Yes, that's system monitoring software, not graphics card monitoring... the cards are both 1GB ATI cards!

I also installed Avalanche, which gave the same readings and finally installed CoreTemp on the advice of a friend. That's quite neat actually - it puts the temperature readout on the LCD screen of my G15 (at last... something useful on there!!).

It shows the same temperatures, and identifies Tj.max as 100C as I recall.

So all three measuring tools show the same numbers. I've heard anecdotal evidence of "sticky" temperature sensors on these chips, but that's not very reassuring.

In fact, I checked when I was home at lunchtime and the front CPU fan is running at a constant 100%, even at idle. I also enabled SpeedStep in the BIOS, but the reported temperatures remain the same regardless of SpeedStep, even with no load (which I would imagine indicates a sensor fault, but I'm way outside my knowledge comfort zone here).

So... yes, something very, very wrong there. I'll make the approrpriate sacrifices to the God of Support and then pick up the phone this evening to see what classic lines they have for me this time around.....

Thanks folks.