[Bug rtl-optimization/122992] speed up loops with TEST_HARD_REG_BIT using EXECUTE_IF_SET_IN_HARD_REG_SET

[mark at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122992

--- Comment #18 from Mark Wielaard <mark at gcc dot gnu.org> ---
(In reply to Sam James from comment #17)
> > On 4 Jun 2026, at 13:27, fche at redhat dot com <gcc-bugzilla@gcc.gnu.org> wrote:
> The modules one is probably bogus as it flips, but the other ones are
> runtime tests so may need avx512. I'll try repro in a few hours as well.

Note that the two test machines use different CPUs, both support avx512, but
different subsets and one is AMD, the other Intel. The x3d1 is a Ryzen 9 7950X
(avx512f avx512dq avx512ifma avx512cd avx512bw avx512vl avx512_bf16 avx512vbmi
avx512_vbmi2 avx512_vnni avx512_bitalg avx512_vpopcntdq), while the sw3bb1 is a
Xeon Platinum 8280 (avx512f avx512dq avx512cd avx512bw avx512vl avx512_vnni)