The IRQ_NOT_LESS... error and the IDE test error is a failing hard drive. That's one issue.
On the memory test, swap the two modules -- if the error follows the memory, it's the module. If it doesn't, it's the board.
Well, errors were reported in both modules,
so if they followed modules, it would still indicate MBor both modules bad.
Or so I thought...
Amazingly, despite having errors reported in both modules,
swapping them did fix the memory errors!
Now if I could only figure how to make stupid OS to skip the bad disk sector.
(Sorry for the delayed response, but I thought I posted one before,
and just now accidently stumbled on this page again)