[Bug rtl-optimization/50190] New: linkpk bench of polyhedron fails during validation with gcc trunk when it is compiled with -Ofast on amd64.

["venkataramanan.kumar.gnu at gmail dot com" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50190

             Bug #: 50190
           Summary: linkpk bench of polyhedron fails during validation
                    with gcc trunk when it is compiled with -Ofast on
                    amd64.
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: venkataramanan.kumar.gnu@gmail.com


Polyhedron benchmark linpk fails during validation when compiled with -Ofast
flag.

(Snip)
> Value= 25.114499300     Target= 23.100000000     Tolerance= 2.0000000000
FAIL <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
> Value=0.27880142600E-10 Target=0.27858826400E-10 Tolerance=0.10000000000E-09
> Value=0.22204460500E-15 Target=0.22204460500E-15 Tolerance=0.10000000000E-14
> Value= 1.0000000000     Target= 1.0000000000     Tolerance=0.10000000000E-07
> Value= 1.0000000000     Target= 1.0000000000     Tolerance=0.10000000000E-07
linpk FAILED    1 fails and    4 passes
(Snip)

exact version of GCC: gcctrunk 1780539
system type: amd64
options given when GCC was configured/built: --disable-bootstrap --without-ppl
--without-cloog --enable-languages=c,c++,fortran 
complete command line that triggers the bug: -Ofast


This issue does not occur when -fprotect-parens is specified with -Ofast.
Current trunk has -fno-protect-parens in -Ofast. So this seems to be a
precision issue as GCC may reorder floating point computations. But I have
reported this to bring it to the notice of community.