https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67937
--- Comment #4 from Joshua Cranmer <Pidgeot18 at gmail dot com> --- (In reply to Richard Biener from comment #3) > Most interesting would be a C testcase that produces the CFG with the bogus > counters ;) Yeah, I know, but doing the minimization on a 5MLOC program takes time.