This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: -mfpmath=sse -mno-80387 troubles
On Fri, 28 Jan 2005 13:50:05 -0800, Richard Henderson <rth@redhat.com> wrote:
> Surprising to me that we're not...
$ make disas|grep ,8
40149e: lea (%edx,%ebx,8),%ebx
4038f6: lea 0x8(,%edi,8),%eax
4043db: lea (%ebx,%edi,8),%eax
40cb38: mov 0x429e90(,%esi,8),%ebx
40cb52: mov 0x429e94(,%esi,8),%ebx
40d0a8: lea (%edx,%ecx,8),%ebx
40d171: lea (%edx,%eax,8),%esi
4112e5: lea 0x0(,%esi,8),%eax
41495f: lea 0x0(,%eax,8),%edx
4168f5: mov 0x14(%ecx,%esi,8),%eax
41693f: lea 0x8(%eax,%esi,8),%edi
416afc: mov 0x14(%ecx,%eax,8),%edx
416b42: mov 0x10(%ecx,%eax,8),%edx
416f9a: mov 0x14(%esi,%ebx,8),%eax
416fbf: mov 0x10(%esi,%ebx,8),%eax
$ make disas|grep ,8|wc -l
15
$ make icc|grep ,8|wc -l
960
Same source.
> Nope. Sample in bugzilla as missed-optimization is welcome.
I'll try to cook something