This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [gfortran testsuite, committed] Re: gfortran.dg/forall_1.f90


On Mon, Jun 06, 2005 at 07:01:38AM +0200, Andreas Jaeger wrote:
> Tobias Schl?ter <tobias.schlueter@physik.uni-muenchen.de> writes:
> 
> F> Andreas Jaeger wrote:
> >> The following tests fail for me on Linux/x86-64 now:
> >> 
> >>    gfortran.dg/forall_1.f90  -O1  execution test
> >>    gfortran.dg/forall_1.f90  -O2  execution test
> >>    gfortran.dg/forall_1.f90  -O3 -fomit-frame-pointer  execution test
> >>    gfortran.dg/forall_1.f90  -O3 -fomit-frame-pointer -funroll-all-loops -finline-functions  execution test
> >>    gfortran.dg/forall_1.f90  -O3 -fomit-frame-pointer -funroll-loops  execution test
> >>    gfortran.dg/forall_1.f90  -O3 -g  execution test
> >> 
> >> Note that the test work correctly compiled as 32-bit binaries - they
> >> just fail as 64-bit tests.  The output is an "Aborted".
> >> 
> > ...
> >> Tobias, any ideas?  Is the code they are testing not 64-bit clean?
> >
> > The test made the assumption that
> >   if (i /= 0) i = 0
> > means that even if i is uninitialized in the beginning it will be zero
> > afterwards.  Since this is the only thing that could be affected by
> > optimization, I've committed this change to the mainline and I will also put
> > this on the 4.0 branch once testing finishes.
> >
> > Does this cure the problem?
> 
> No, it does not for me,

forall_1.f90 fails on FreeBSD even without optimization.  The
failure is with the third "call abort", but I haven't been
able to deduce the problem.

-- 
Steve


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]