Re: AMBER: Intel Fortran Compiler Problem

From: Robert Duke <rduke.email.unc.edu>
Date: Wed, 15 Sep 2004 11:38:03 -0400

Oops. The below advice is for amber 8; I lost track of the fact we are
dealing with amber 7 here. On looking back at the amber 7 stuff, most of
the machine files don't even have LOADCC defined, but it IS used in the
ptraj makefile. Now the story here might be it would normally be picked up
in the environment of the make, but it is not in your case, or is wrongly
defined. If I were you I would try sticking a:
setenv LOADCC gcc
in your machine file and seeing what happens.
Regards - Bob Duke

----- Original Message -----
From: "Robert Duke" <rduke.email.unc.edu>
To: <amber.scripps.edu>
Sent: Wednesday, September 15, 2004 11:30 AM
Subject: Re: AMBER: Intel Fortran Compiler Problem


> Hmmm. I am not at ptraj expert. However, I note that LOADCC here is
> mapping to ifort, whereas it would normally map to gcc, even if you are
> using ifort. Did you maybe whack config.h, changing this value?. You are
> having trouble with two "main" routines, one from rdparm.c, and one from
the
> fortran libraries. You want the one from rdparm.c, which was
gcc-compiled,
> presumably, and when linked with gcc, gcc will use the main() in rdparm.c.
> This is all guesswork; I happen to compile the ptraj project (which
includes
> rdparm) with g77 currently (no reason other than convenience).
> Regards - Bob Duke
>
> ----- Original Message -----
> From: "Stephen P. Molnar, Ph.D." <s.molnar.sbcglobal.net>
> To: "Amber" <amber.scripps.edu>
> Sent: Wednesday, September 15, 2004 11:38 AM
> Subject: Re: AMBER: Intel Fortran Compiler Problem
>
>
> > Thanks for the reply. That took care of the first error.
> >
> > I got quite a bit further before:
> >
> > ar cru libpdb.a pdbrun.o pdb_read.o pdb_sprntf.o pdb_sscanf.o
> > pdb_write.o ms.o
> > : libpdb.a
> > make[2]: Leaving directory `/home/computation/amber7/src/ptraj/pdb'
> > ../Compile LOADCC -o rdparm main.o rdparm.o dispatch.o help.o utility.o
> > io.o trajectory.o torsion.o rms.o display.o interface.o energy.o
> > experimental.o ptraj.o actions.o analyze.o pdb/libpdb.a -lm
> > /opt/intel_fc_80/bin/ifort -o rdparm main.o rdparm.o dispatch.o help.o
> > utility.o io.o trajectory.o torsion.o rms.o display.o interface.o
> > energy.o experimental.o ptraj.o actions.o analyze.o pdb/libpdb.a -lm
> > main.o(.text+0x68): In function `main':
> > : multiple definition of `main'
> > /opt/intel_fc_80/lib/for_main.o(.text+0x0): first defined here
> > ld: Warning: size of symbol `main' changed from 42 in
> > /opt/intel_fc_80/lib/for_main.o to 271 in main.o
> > /opt/intel_fc_80/lib/for_main.o(.text+0x1c): In function `main':
> > : undefined reference to `MAIN__'
> > make[1]: *** [rdparm] Error 1
> > make[1]: Leaving directory `/home/computation/amber7/src/ptraj'
> > make: *** [install] Error 2
> >
> >
> > On Wed, 2004-09-15 at 10:52, Robert Duke wrote:
> > > Maybe leave LOADLIB defined as nothing in the MACHINE file?
> > > ----- Original Message -----
> > > From: "Stephen P. Molnar, Ph.D." <s.molnar.sbcglobal.net>
> > > To: "Amber" <amber.scripps.edu>
> > > Sent: Wednesday, September 15, 2004 11:13 AM
> > > Subject: Re: AMBER: Intel Fortran Compiler Problem
> > >
> > >
> > > > Thanks for your reply.
> > > >
> > > > On Wed, 2004-09-15 at 10:33, Viktor Hornak wrote:
> > > > > Stephen P. Molnar, Ph.D. wrote:
> > > > > > I am attempting to compile Amber v7 for my SuSE9.0 linux
partition
> > > using
> > > > > > the Inter Fortran Compiler v8.
> > > > > >
> > > > > > Unfortunately, I got the following error early on:
> > > > > >
> > > > > > /opt/intel_fc_80/bin/ifort -o new2oldparm new2oldparm.o nxtsec.o
> > > > > > -lPEPCF90
> > > > > > ld: cannot find -lPEPCF90
> > > > > >
> > > > > > Now, according, to the locate command PEPCF90 is part of the v7
> Intel
> > > > > > compiler, and is not to be found in v8.
> > > > > >
> > > > > > What is the solution to this problem.
> > > > >
> > > > > removing the -lPEPCF90 from the MACHINE file should do it...
> > > > >
> > > > Good suggestion! Unfortunately, it generated a new error:
> > > > mkdir ../exe
> > > > cd lib; make install
> > > > make[1]: Entering directory `/home/computation/amber7/src/lib'
> > > > ../Compile L0 -P -DNEWPARM nxtsec.f
> > > > cat nxtsec.f | /lib/cpp -traditional -P -DNEWPARM -DLinux
> > > > -DMEM_ALLOC > _nxtsec_.f
> > > > /opt/intel_fc_80/bin/ifort -c -w -O2 _nxtsec_.f
> > > > ../Compile LOAD -o new2oldparm new2oldparm.o nxtsec.o
> > > > LOADLIB: Undefined variable.
> > > > make[1]: *** [new2oldparm] Error 1
> > > > make[1]: Leaving directory `/home/computation/amber7/src/lib'
> > > > make: *** [install] Error 2
> > > >
> > > >
> > > > > Cheers,
> > > > > -Viktor
> > > > --
> > > > Stephen P. Molnar, Ph.D. Life is a fuzzy set
> > > > Foundation for Chemistry Stochastic and multibariant
> > > > http://www.geocities.com/FoundationForChemistry
> > > >
> > >
> > -----------------------------------------------------------------------
> > > > The AMBER Mail Reflector
> > > > To post, send mail to amber.scripps.edu
> > > > To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
> > > >
> > >
> > >
> >
> -----------------------------------------------------------------------
> > > The AMBER Mail Reflector
> > > To post, send mail to amber.scripps.edu
> > > To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
> >
> > -----------------------------------------------------------------------
> > The AMBER Mail Reflector
> > To post, send mail to amber.scripps.edu
> > To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
> >
>
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>


-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Sep 15 2004 - 16:53:01 PDT
Custom Search