My own opinion is that mpich2 should have been put in /opt/mpich2 or somewhere out of /usr at the very least.
I have no idea if Apple creates parallel routines which depend on openmpi in /usr, so I leave that directory alone.
I also put /opt/mpich2/bin at the head of my path to make sure I'm using the mpich2 wrappers:
-- example --
PATH=$AMBERHOME/exe:/opt/mpich2/bin:/usr/local/CPMD-3.13.2/bin:/usr/local/gamess:/Library/MGLTools/1.5.2/bin:/opt/local/bin:/opt/local/sbin:/usr/local/bin/sdftoolkit:/Applications/moe/bin:$PATH
export PATH
-- end example --
Perhaps George also needs to set MPI_HOME.
In particular, it appears that something is very wrong with this line:
gfortran -I/usr/local/include -I/usr/local/lib -L/usr/local/lib -lmpi_f90 -lmpi_f77 -lmpi -lopen-rte -lopen-pal -lutil
Those are openmpi libs, not mpich2 libs.
On May 31, 2010, at 1:02 PM, case wrote:
> On Mon, May 31, 2010, George Tzotzos wrote:
>
>> I did "nm schlegel_dg.o | grep schlegel". Below's the output. It seems
>> that I have treble underscore.
>
> Don't worry about those.
>
>>
>> $ nm schlegel_dg.o | grep schlegel
> ...
>> 0000000000000000 T _schlegel_dg_
>
> This shows that you should have an entry point "_schlegel_dg_" in the object
> file. So I'm lost as to why the load step doesn't seem to be able to find it.
> Maybe someone else on the list sees what I am missing. It would still help
> to see the full error message, not just the snippets you had in the previous
> post.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon May 31 2010 - 11:31:05 PDT
