) (12/13/89)
has anyone noticed differences in floating-pt behavior between Exp-1 and Exp-2? I have some complex codes which depend on floating-pt values, and I get different answers between the two CPUs (indeed, with the Exp-2, sometimes I don't get an answer at all--function-calls never return). any ideas? this sounds like a microcode bug, but I'm not sure to to isolate it well enough to be certain. -- clint