KA620

From: Paul Koning <pkoning_at_equallogic.com>
Date: Tue Mar 23 08:08:53 2004

>>>>> "Antonio" == Antonio Carlini <arcarlini_at_iee.org> writes:

 Antonio> I never had access to AXE, but as I understand it, it would
 Antonio> generate semi-random sequences of instructions and then
 Antonio> execute them. Then it would compare the actual results with
 Antonio> known good results. I don't know the details of how it would
 Antonio> determine "known good results" given that the sequence was
 Antonio> semi-random.

I assume the answer is "by running the random instructions in a
simulation of what the SRM says".

Incidentally, it turns out this is not a new technique. CDC had one
(or two?) random instruction generator diagnostics for the 6600
supercomputer, back in the mid to late 1960s. Those were actually
used in the field, which makes sense given that those machines had
CPUs built out of hundreds of small (discrete transistor) modules all
wired together by twisted pair wire.

      paul
Received on Tue Mar 23 2004 - 08:08:53 GMT