https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127610
--- Comment #14 from Drea Pinski <pinskia at gcc dot gnu.org> --- So only the last one which seems different and only slightly: r17-4132-ga8c97564fbc39b: tree VRP : 0.23 ( 11%) 7885k ( 7%) dominator optimization : 0.25 ( 12%) 10M ( 9%) backwards jump threading : 0.19 ( 9%) 24M ( 22%) tree CFG cleanup : 0.05 ( 2%) 284k ( 0%) df live regs : 0.13 ( 6%) 0 ( 0%) r17-4133-g230d950810f378: tree VRP : 0.34 ( 10%) 11M ( 7%) dominator optimization : 0.35 ( 10%) 15M ( 10%) backwards jump threading : 0.29 ( 8%) 32M ( 21%) tree CFG cleanup : 0.08 ( 2%) 621k ( 0%) df live regs : 0.26 ( 7%) 0 ( 0%) Which tells me that the change has indirect effect on other optimization passes. VRP memory usage increased from 7M to 11M which points to it doing more work.
