[Bug tree-optimization/84997] Optimize integer operations on floating point constants without -ffast-math but with -fno-trapping-math

[pinskia at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84997

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Optimize integer operations |Optimize integer operations
                   |on floating point constants |on floating point constants
                   |without -ffast-math         |without -ffast-math but
                   |                            |with -fno-trapping-math

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
>clang optimizes to the following assembly

Note clang defaults to -fno-trapping-math.
Using -ftrapping-math with clang produces what GCC produces currently:
        cvtsi2sd        %edi, %xmm0
        addsd   .LCPI0_0(%rip), %xmm0
        cvttsd2si       %xmm0, %eax
        retq