https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92508
--- Comment #13 from Martin Liška <marxin at gcc dot gnu.org> --- (In reply to Jan Hubicka from comment #12) > > It still fails for me with the current trunk: > > > > $ ../configure --prefix=/tmp/gcc/ --disable-multilib --without-isl > > --disable-libsanitizer --with-build-config=bootstrap-lto-lean > > $ make profiledbootstrap -j123 > > I did profiledbootstrap thru lunch and except that I hit some -Werror > issues it passed w/o problems. I will try to replicate it with your > flags. You may look how much the values differ if it looks like another > roundoff or something else. Did you use bootstrap-lto-lean (and not bootstrap-lto)? > The test should disable itself for profiled functions since for IPA > count we subtract the time spent in inline clones from the main function > which is prone to roundoff errors so one can't test time to match > precisely. > > Honza
