https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97983
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2020-11-25
Status|UNCONFIRMED |NEW
Ever confirmed|0 |1
--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Ah, the profiledbootstrap issue is just that I had the compiler configured
--enable-checking=release. With normal bootstrap, stage2 is built with
-fchecking=1 and that ICEs, while in profiledbootstrap stagetrain is not built
with that and results with -da in 1 GB *.reload dump and this extremely huge
stack frame for the function. I bet if the checking issue is fixed the
non-checking problem will go away too.