[Bug target/120941] [16 Regression] 24-40% slowdown of 519.lbm_r on Zen2 and 470.lbm on Zen5 since r16-1644-gaba3b9d3a48a07

[pheeck at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120941

--- Comment #43 from Filip Kastl <pheeck at gcc dot gnu.org> ---
(In reply to H.J. Lu from comment #42)
> Created attachment 62020 [details]
> A new patch
> 
> Here is a patch not to limit non all 0s/1s vector loads in the same loop.
> Please try it.

This patch also helps get the exec time to the original ~163s.