https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86462
--- Comment #1 from rguenther at suse dot de <rguenther at suse dot de> --- On July 10, 2018 5:15:08 PM GMT+02:00, "marxin at gcc dot gnu.org" <gcc-bugzi...@gcc.gnu.org> wrote: >https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86462 > > Bug ID: 86462 > Summary: [9 Regression] Quite huge debug info size increase > introduced in r262511 > Product: gcc > Version: 9.0 > Status: UNCONFIRMED > Severity: normal > Priority: P3 > Component: debug > Assignee: unassigned at gcc dot gnu.org > Reporter: marxin at gcc dot gnu.org > CC: rguenth at gcc dot gnu.org > Target Milestone: --- > >Buidling induct2.f90 from polyhedron benchmark: >https://www.polyhedron.com/web_images/documents/pb11.zip > >$ gfortran induct2.f90 -O2 -g -c && du -hs induct2.o > >$ bloaty bad.o -- good.o > VM SIZE FILE SIZE > ++++++++++++++ GROWING ++++++++++++++ > [ = ] 0 .rela.debug_ranges +84.5Ki +21e2% > [ = ] 0 .rela.debug_info +22.2Ki +24% > [ = ] 0 .debug_info +1.02Ki +7.5% > [ = ] 0 .debug_ranges +36 +78% > [ = ] 0 .debug_abbrev +16 +2.5% > > [ = ] 0 TOTAL +107Ki +24% Interesting. This either means we lost info before or we are outputting scopes early that we do not need.