https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120614
--- Comment #4 from kugan at gcc dot gnu.org --- x264_pixel_sad_x4_16x16.diff is at -O3 without -flto. Function level profiling is same even with -flto. x264_pixel_sad_x4_16x16 total:18508 head:4627 0: 4627 0.1: 0 0.2: 0 0.3: 0 0.4: 0 0: x264_pixel_sad_16x16 total:13881 0.4: 4627 0.5: 4627 0.6: 4627 0.8: 0 Branch probably with the AutoFDO is vastly different.