In your arch.make, did you specify mpif90 as your FC and also, did you set the following block?
MPI_INTERFACE=libmpi_f90.a MPI_LIBS= DEFS_MPI=-DMPI MPI_INCLUDE=/opt/openmpi/gnu/include <----------Change this to your MPI choice Daniel J. Backlund Texas Tech University - Physics [email protected] ________________________________________ From: Hegoi Manzano [[email protected]] Sent: Monday, March 01, 2010 9:24 AM To: [email protected] Subject: [SIESTA-L] FW: Parallel compilation, version 2.0.2 Hi, I have compiled successfully Siesta in an intel_64x machine. The program runs without problems in serial, but I cannot run it in parallel. When I run mpirun -np 4 ./siesta < input> output Instead running the code in 4 nodes, it runs 4 times the same calculation in one node. I guess that this is because I didn't compile it properly to work in parallel. Can you give me any clue to modify this? Best _____________________________________________ Dr. Hegoi Manzano Concrete Sustainability Hub. Massachusetts Institute of Technology, Room 1-370 77 Massachusetts Avenue, Cambridge, MA02139 e-mail: [email protected] // [email protected] www.hegoi-manzano.blogspot.com _________________________________________________
