https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85103
--- Comment #17 from Jan Hubicka <hubicka at gcc dot gnu.org> --- I am re-doing benchmarks now to see where we are standing with gcc9. I have checked reducing max-inline-insns-single as Richard mentioned, reducing to 200 or 300 basically brings one regression and that is CactusADM, about 29%. I will experiment with the hints and see if I can tweak the metrics here.
