Every time my code has done that, it was an out of bounds index into a
floating point array. If you're sure your code generates really small
numbers (iterative code can underflow pretty easily), you could add
a signal handler to patch things up. I've seen it done on an RS/6000
but haven't seen it done on Linux. I read somewhere that Alpha by
default traps ALL floating point exceptions. Also read it was possible
to turn that off. Think it was in one of the Alpha manuals you can
get from DEC's web site. As I mentioned, I just corrected the code
for the stuff I've been working on. I do have an iterative code that
I need to get working but haven't gotten to it yet. So, if you figure
out how to turn it off please post the solution....
Wes