[Bug tree-optimization/126551] missed vectorization due to missed reduction detection in ifcvt in miniBUDE

"burnus at gcc dot gnu.org via Gcc-bugs" <[email protected]> Wed, 05 Aug 2026 11:13:15 +0000

Newsgroups	gmane.comp.gcc.bugs
Message-ID	<[email protected]/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126551

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|x86_64-linux-gnu            |x86_64-linux-gnu,aarch64-un
                   |                            |known-linux-gnu
                 CC|                            |burnus at gcc dot gnu.org

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Likewise for aarch64-unknown-linux-gnu

With (and only with) the manual change for combining the two 'etot +=' (see
comment 0), the result is (-Ofast -fopt-info-vec-optimized):

<source>:91:22: optimized: loop vectorized using 16 byte vectors and unroll
factor 4