Yes, I'm seeing this problem on one of carlos's machines (I'm helping get
the NEB code debugged and committed). I had to add the -L flag to the FLIBS
variable to get compilation to work. I didn't see this before, but since I
havent really had an intel machine to play with I'm not sure what changed
and when. This doesn't seem to bother my gnu/cygwin setup.
I tried using ifort to link and that gave me a new set of errors.
On Thu, Mar 4, 2010 at 9:10 AM, case <case.biomaps.rutgers.edu> wrote:
> On Thu, Mar 04, 2010, Daniel Roe wrote:
>
> > This seems to be happening because icc can't find the fortran routines it
> > needs. Adding a -L flag with the directory to your fortran library
> > directory, something like e.g.
> >
> > icc -o ptraj main.o rdparm.o dispatch.o help.o utility.o second.o io.o
> > trajectory.o netcdf_ptraj.o parallel_ptraj.al.o ptraj.o actions.o
> analyze.o
> > thermo.o pubfft.o cluster.o clusterLib.o* -L/opt/intel/fce/10.1.015/lib*
> ...
> >
> > before all of the libraries seems to do the trick. I'm not sure what the
> > best way to correct this in the Makefile/configure script is.
>
> Dan: do you have this problem on *your* machine (before you make the "-L"
> fix?) Is the icc compiler finding some other libicore.a file? I'm unsure
> how icc knows how to find the ifort libraries in the older compilers (pre
> 11)
> where icc and ifort installations were in different directory trees.
>
> We could just link with FC rather than CC, and it's worth a try to see if
> that
> works. But it will be a job to make such a change to nab.
>
> ...thx...dac
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Thu Mar 04 2010 - 06:30:03 PST
