[Bug target/117126] [14/15/16 regression] GCC 14 generates redundant movq xmm1, xmm1 instructions with -O2 optimization, leading to code bloat (regression from GCC 13.x)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117126

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |13.4.0
     Ever confirmed|0                           |1
           Priority|P3                          |P2
   Last reconfirmed|                            |2026-02-16
             Status|UNCONFIRMED                 |NEW

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  IIRC the way we do half-SSE vectors was rewritten because of some
correctness corner case, the GCC 13 emitted code looks correct and is indeed
better.