[Bug tree-optimization/71815] SLSR misses several PHI candidate cases

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71815

--- Comment #9 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
I'm not comfortable with the results of the patch.  Overall I see a slight
improvement for SPECint CPU2006 and a slightly larger degradation for SPECfp
CPU2006.  But there are some individual slowdowns that are unacceptable:

435.gromacs:   -8.7%
436.cactusADM: -3.0%

On the other hand:

403.gcc:       +4.6%

All other results within +/- 2%.

I'm going to hold off on this until I can investigate the 435.gromacs slowdown.