On Jan 9, 2009, at 3:08 AM, Paolo Bonzini wrote:
Is it at all possible to get the x86 FPU to produce the correct result for this situation?
You can also try compiling the Squeak VM with -mfpmath=sse.
That should work, as long as you are running it on x86 processors that support SSE2 with double-precision scalar operations. The SSE floating-point unit(s) perform true single- and double-precision arithmetic.
-- tim