https://gcc.gnu.org/bugzilla/show_bug.cgi?id=41647
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |RESOLVED
CC| |matz at gcc dot gnu.org
Resolution|--- |WONTFIX
--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
I don't think early unrolling "complicates" polyhedral analysis anymore.
There's some obvious issues with the GRAPHITE implementation with DRs being
globbed at
basic-block granularity being the most obvious one. Micha has patches^Whacks
for some of these.
Anyways I believe this particular issue (should not be) is not an issue.
Generally early complete unrolling is important to remove abstraction penalty.
