This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran 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 44 optimizations and Intel SSE4


Using "-O2" also resulted in the same link errors. 

-- jake


> -----Original Message-----
> From: Steve Kargl [mailto:sgk@troutmask.apl.washington.edu]
> Sent: Tuesday, November 01, 2011 3:10 PM
> To: Jacob Anderson
> Cc: fortran@gcc.gnu.org
> Subject: Re: GFortran 44 optimizations and Intel SSE4
> 
> On Tue, Nov 01, 2011 at 03:02:27PM -0700, Jacob Anderson wrote:
> > Well, we are using "-lstd++" in our link directive - this is a mixed
> > c++/C/F90 program which is built on MPI. The mpif90 script is used to
> link
> > the entire project together.
> >
> > There is something amiss with the "-O3" option though. It should not
> change
> > the link image and cause incompatibility with the std++ library on
> platform.
> > We DO have both a 4.1 and 4.4 version of G(cc|fortran) installed.
> Maybe the
> > "-O3" is pulling in the 4.1 version of the library.
> >
> > In /usr/lib there is an so.8 and so.5 version of libstd++.
> >
> > Without the "-O3" option, as mentioned in the other emails, works just
> fine
> > - full link and successful production of binary.
> >
> 
> Well, there's always -v that spews out the underlying
> details of the compile and link stages.
> 
> Something seems to be messed up on your system,
> because -O3 should have no effect on pulling in or
> discarding -lstdc++ (unless the mpif90 wrapper is
> doing something dogdy).
> 
> --
> Steve



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