Re: [xine-user] CFLAGS breaking builds and fucking up optimalisations

[Graham M. <gr...@gm...>]

Chris Chabot <ch...@xs...> writes:

> Ps you could consider using -march=i686, thats about the most
> optimisation gcc 3.x can do (p3/4/athlon is max 1% diff in code
> speed). that way you do provide optimized code, without trigering
> internal compiler bugs ;-)

It looks to me as though the configure script has the tests in the
wrong order. It tests for '-cpu=' first and only if that fails does
it try '-march='.