better optimization flags for opteron (amd64)// -ffast-math and -O3 are not friends

[Olexandr Konovalenko]

hi!

I found that on my system combination of -ffast-math AND -O3 is slower than
-O0 on AMD64 (interestingly it is true for Xeon EM64 as well), how can it be?
though just -O3 gives fastest double precision floating point computations.
On  32bit platform this effect is not observed.

please look here: http://daemon.nanophys.kth.se/~kono/testfcpu

the compiler I use:
Using built-in specs.
Configured with: FreeBSD/amd64 system compiler
Thread model: posix
gcc version 3.4.2 [FreeBSD] 20040728

my platform:
uname -a
FreeBSD daemon.nanophys.kth.se 5.4-STABLE FreeBSD 5.4-STABLE #0: Tue Apr 19
20:26:31 CEST 2005
root at daemon.nanophys.kth.se:/usr/obj/usr/src/sys/DAEMON64  amd64

-- Alexander Konovalenko