honestly it could be quite a few things, so lets try and narrow it down.
heat.
Code:
cat /proc/acpi/thermal_zone/THRM/temperature
do that every few minutes to see if the temperature finds a balance, or wheter it just climbs.
daemon crashing, kernel crashing
read the entries in /var/log/messages
be sure the check them in this order, becuase if it is overheating, more than likely you will see a kernel panic or strange error