https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123587
--- Comment #12 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #11) > I am still trying to reduce a testcase . Still reducing. I was able to remove 2 files (including the C source). I also starting using a makefile to build so building the lto object files were in parallel rather than serial; reduces the failure time in 1/4 so it is faster to reduce. It is still slow. I did some manual removal of code too.
