[Bug rtl-optimization/119285] [15 Regression] 5% slowdown of 519.lbm_r on Zen2 and Zen4 since r15-7932-ge355fe414aa3aa

[vmakarov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119285

--- Comment #1 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
I can not reproduce it on Intel (13600K).  GCC with patch and without the patch
has the same score for lbm_s (11.4).  But I see big lbm_s code increase with
the patch (+0.87%).  And this is suspenseful.

In any case I will investigate what is changed in the code with the patch.  If
I have a solution I will run spec2017 on intel and a zen machine to check that
there is no significant performance change.  Probably I will find a solution on
the next week.