Christoph Groth wrote:
If I take the second (-O1) binary and execute it on an Opteron it produces the _first_ result. Thus, the same static binary produces different results on different processors!
That is not unusual. The 64-bits only give a precision of about 15-16 decimal places, so you can't necessarily expect identical results on different processors.
If you want that, then look at a software solution for the floating point maths - something like mpfr. However, I know from experience that for Monte Carlo simulations, they often need a lot of CPU time, which would rise dramatically if you swapped from hardware to software floating point.
Dave
