https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125871
H.J. Lu <hjl.tools at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2026-06-18
Ever confirmed|0 |1
--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Tamar Christina from comment #2)
> (In reply to Richard Biener from comment #1)
> > I'll note that
> >
> > /* { dg-additional-options "-mavx512f" { target { i?86-*-* x86_64-*-* } } }
> > */
> > /* { dg-require-effective-target avx512f_runtime { target i?86-*-*
> > x86_64-*-* } } */
> >
> > is overly restrictive and should do
> >
> > /* { dg-additional-options "-mavx512f" { target avx512f_runtime } } */
> >
> > And gate the dump scans on { target avx512f }
> >
> > I'll also note that we should really fixup check_vect () to gate AVX512
> > enabled features ...
>
> Yeah I don't really understand all the knobs so I copy things from other
> tests.
>
> I build x86 with -march=native because this gives more vectorization. But
> this makes it impossible to whether I need random flag to make things pass
> when not.
>
It is independent of -march=native. Please try
make check-gcc RUNTESTFLAGS="--target_board='unix{-m32,}'
vect.exp=gcc.dg/vect/vect-early-break_144-pr125804.c"
It fails with -m32.