[Bug middle-end/42298] [4.5 Regression] GCC fails validity in mdbx in polyhedral benchmark.

[howarth at nitro dot med dot uc dot edu]

------- Comment #4 from howarth at nitro dot med dot uc dot edu  2009-12-07 14:03 -------
(In reply to comment #2)
> Huh?  Does plain -O2 work?  Do I understand correctly that -O2 -fno-loop-block
> -fno-loop-strip-mine miscompiles?
> 
> Sebastian, how can disabling graphite options but not enabling graphite
> miscompile anything??  P1, please fix / comment.
> 

Richard,
    Although this doesn't appear to be the case here (because the reporter
claims -O2 also fails), one could have graphite cause miscompiles without other
graphite options. The -graphite-identity option just converts the GIMPLE to the
polyhedral representation and back to GIMPLE. This step could go wrong as is
seen in PR42181 with the air.f90 benchmark.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42298