https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84490
Bug ID: 84490 Summary: 436.cactusADM regressed by 6-8% percent with -Ofast on Zen, compared to gcc 7.2 Product: gcc Version: 8.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: target Assignee: unassigned at gcc dot gnu.org Reporter: jamborm at gcc dot gnu.org Target Milestone: --- Compared to gcc 7.2, 436.cactusADM is about 8% and 6% slower when run on an AND Ryzen and an AMD EPYC respectively after compiling it on trunk with -Ofast (and generic march and tuning). So far I have not done any bisecting, only verified that this is most probably unrelated to PR 82362 because r251713 does not seem to have any effect on the benchmark (the benchmark is already slow with that revision).