https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69196
--- Comment #26 from Jeffrey A. Law <law at redhat dot com> --- I'm planning to keep it open until I add the code to factor FSM path clones with differing entry edges into a single FSM path clone. At that time we'll want to look at this (and other) BZs to ensure that we've got the tuning nailed down.
