ULSI vs Cyrix 387's

[Samuel A. Figueroa]

There is a thread right now on comp.os.os2.misc on a similar subject: a program
which performs f-p division by zero "crashes" the OS/2 operating system.  It's
not yet clear to me if this is the case for only certain compilers, though it
certainly happens with a particular version of Microsoft's FORTRAN compiler.
The claim is that this behavior is not dependent on the compiler used.  So far,
no one has said he or she has this problem with an Intel math coprocessor,
though some 486 motherboards have this problem apparently due to poor design.
This problem has been reported with IIT and Cyrix math coprocessors.

Sam