[Bug rtl-optimization/98782] [11/12 Regression] Bad interaction between IPA frequences and IRA resulting in spills due to changes in BB frequencies

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98782

--- Comment #30 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
(In reply to Martin Jambor from comment #29)
> (In reply to Tamar Christina from comment #23)
> > I wonder if we can get rid of the final magic parameter too, we run with
> > --param ipa-cp-unit-growth=80 too which seems to have no more effect on
> > exchange, though still a slight effect on leela but a 12% gain in imagick.
> > 
> > This looks like without the parameter we lose constprop on MorphologyApply
> > resulting in a much bigger function.  Do you want me to make a new ticket
> > for this?
> 
> Indeed I did not know about this.  I think tracking it in a (separate) PR
> would make sense.  Thanks a lot for the pointer!

Thanks, created a new ticket
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103734 the threshold seems to be a
lot lower than 80%, 20% seems to already be enough.