https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117932
--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Richard Biener from comment #1) > With -fno-ext-dce -fno-fold-mem-offsets on trunk the identified regressions > distribute to > > if-conversion 2 : 3.98 ( 2%) 55k ( 0%) > `- dominance computation : 1.02 ( 0%) 0 ( 0%) > `- loop init : 0.20 ( 0%) 1512 ( 0%) > `- df live regs : 17.28 ( 7%) 0 ( 0%) > `- df live&initialized regs : 15.76 ( 6%) 0 ( 0%) > > hard reg cprop : 4.05 ( 2%) 9072 ( 0%) > `- df reg dead/unused notes : 0.11 ( 0%) 278k ( 0%) > `- df live regs : 11.19 ( 4%) 0 ( 0%) > `- df live&initialized regs : 10.37 ( 4%) 0 ( 0%) > > reorder blocks : 13.82 ( 5%) 4663M ( 80%) > `- rebuild jump labels : 0.04 ( 0%) 0 ( 0%) > `- cfg cleanup : 1.41 ( 1%) 904 ( 0%) For aarch64 it seems similar (--enable-checking=yes but -fno-checking): if-conversion 2 : 29.43 ( 1%) 10k ( 0%) `- df live regs : 113.36 ( 4%) 0 ( 0%) `- loop init : 4.79 ( 0%) 1512 ( 0%) `- dominance computation : 12.42 ( 0%) 0 ( 0%) `- df live&initialized regs : 53.90 ( 2%) 0 ( 0%) hard reg cprop : 44.58 ( 2%) 10k ( 0%) `- df live regs : 106.39 ( 4%) 0 ( 0%) `- df live&initialized regs : 53.82 ( 2%) 0 ( 0%) `- df reg dead/unused notes : 0.35 ( 0%) 45k ( 0%) reorder blocks : 90.75 ( 4%) 4664M ( 77%) `- cfg cleanup : 27.61 ( 1%) 136 ( 0%) `- rebuild jump labels : 0.11 ( 0%) 0 ( 0%) scheduling fusion : 19.70 ( 1%) 1551k ( 0%) `- df reg dead/unused notes : 0.11 ( 0%) 0 ( 0%) `- alias analysis : 5.15 ( 0%) 2048k ( 0%) `- df live&initialized regs : 27.27 ( 1%) 0 ( 0%) `- register information : 0.06 ( 0%) 0 ( 0%) `- df live regs : 60.21 ( 2%) 0 ( 0%) scheduling 2 : 20.40 ( 1%) 1753k ( 0%) `- df reg dead/unused notes : 0.15 ( 0%) 0 ( 0%) `- alias analysis : 4.93 ( 0%) 2048k ( 0%) `- df live&initialized regs : 27.61 ( 1%) 0 ( 0%) `- register information : 0.10 ( 0%) 0 ( 0%) `- df live regs : 61.71 ( 2%) 0 ( 0%)
