I've been having this problem since I first built my system a few weeks back. When I'm doing something intensive like gaming, gpu stress test or cpu stress test - actually those are the only things to cause it so far - my system will automatically reboot, but it takes a minute or so between shutting down and actually booting up again. My measurable temps are a little high but definitely still safe and would not cause a thermal shutdown. Both my CPU and graphics card stay at around 75c under full load, though the CPU is only ever that during the stress test. My PSU is only a little warm to the touch when it happens too, nothing out of the ordinary.
I think I have it narrowed it down to three possible things (could I be missing anything?):
Chipset overheating which I can't monitor
PSU over or undervolting components
I figure I can only find out if it's the mobo through process of elimination, the chipset I'll know more about tomorrow when my new fans arrive and I improve my cooling, so that just leaves me with the PSU for now and I need a way of logging voltages. I have HWMonitorPro but it doesn't save logs if the systems shuts down unexpectedly. Keeping my eye on it can't really be done either because it'll only be a split second before it shuts down. Are there any other options that don't involve a multimeter?
Thanks for the reply. I took out one stick of RAM and during a GPU stress test it restarted as usual, but then I swapped that stick with the one I previously removed and now something very strange is happening. I ran the Kombustor again but this time it seemed to work fine, right up until I pressed ESC to exit full screen which cause it to restart. It has done this 4 times in a row now. I don't know what to make of it.