https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127203
--- Comment #5 from Filip Kastl <pheeck at gcc dot gnu.org> --- Looks like that helped a lot. I see many graphs back on the original numbers. There are still graphs with ~10% slowdown: AMD Zen3 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=750.591.8 AMD Zen2 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=807.591.8 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=811.591.8 Intel Ice Lake (3rd generation Xeon) https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=691.591.8 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=694.591.8 AMD Zen4 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=696.591.8 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=698.591.8 AMD Zen5 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=991.591.8 https://lnt.opensuse.org/db_default/v4/CPP/graph?plot.0=981.591.8 I'm okay with both closing the bug and leaving it open.
