SSE Optimization control

[corey taylor]

Hi guys,

  We've been trying to find a way to write code that uses SSE and has
a fallback when not supported on the system currently running.  This
would seem simple enough -- check the cpuid and branch accordingly.

  However, when enabling -msse or -msse2 in gcc, integer/floating
point cast optimizations are inserted everywhere and causes an
immediate SIGILL of course on an incompatible system.

  The SSE code is written in assembly.  Is there away to control these
auto-optimizations and do we need to enable it for hand-written SSE?

Corey