[Bug rtl-optimization/81611] [8 Regression] gcc un-learned loop / post-increment optimization

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81611

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |4.7.3
   Target Milestone|---                         |8.0
            Summary|gcc un-learned loop /       |[8 Regression] gcc
                   |post-increment optimization |un-learned loop /
                   |                            |post-increment optimization
      Known to fail|                            |8.0

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Looks like a register allocation / SSA coalescing issue to me.  The IV is
either not coalesced on the backedge or RA splits the live range.  Maybe IVOPTs
related as well.

I suspect unless you do some more digging (bisection?) nothing will happen here
;)