https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87223

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization, ra
             Target|                            |x86_64-*-* i?86-*-*
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2018-09-05
                 CC|                            |vmakarov at gcc dot gnu.org
          Component|c++                         |target
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  I believe there are duplicates for this and I believe this
optimization would be beneficial for -O2+ as well.  Note that with -O2 we
generate

        xorl    %eax, %eax
        movb    $0, (%rdi)
        movq    $0, 8(%rdi)
        movw    %ax, 16(%rdi)
        movq    $0, 24(%rdi)
        ret

the movw variant is generated by peephole2.

I suppose LRA could be the place to "remat" constants.

Reply via email to