This guide was created for use with LAM-MPI version 7.1.4 using the PGI 7.1 Release. This information is for both x64 processors running 64-bit Linux or x86 processors running 32-bit Linux.
LAM-MPI is a freely available implementation of MPI. Information about LAM-MPI can be found at the LAM-MPI home page
LAM-MPI is freely available for download from the LAM-MPI download page.
There are no known dependencies.
After you have downloaded and unpacked the LAM-MPI packages, the typical build process is followed. After setting up the environment and determining the final installation directory, you run configure followed by make and finally make install Please modify this example as needed. Be sure to change the prefix path. We assume the PGI compilers are installed in the default /opt/pgi directory.
cd lam-7.1.4 export CC=pgcc export CXX=pgCC export F90=pgf90 export FC=pgf77 export CFLAGS=-D__GNUC__ ! only needed for 64-bit export CCFLAGS=-D__GNUC__ ! only needed for 64-bit ./configure --prefix=/opt/pgi/linux86/7.1/mpi/lam ! for 32-bit ./configure --prefix=/opt/pgi/linux86-64/7.1/mpi/lam ! for 64-bit
To build and install, simply run 'make' and 'make install'.
make make install
After you have successfully installed the LAM-MPI library you will need to create a 'machine' file with the list of your cluster's host names. See the LAM-MPI User's Guide
To test your LAM-MPI installation, the LAM-MPI package contains several example test programs. First make sure that the LAM-MPI install bin directory is in your PATH environment variable. Then build and run the examples.
export PATH=/opt/pgi/linux86/7.1/mpi/lam/bin /opt/pgi/linux86/7.1/bin:$PATH ! for 32-bit export PATH=/opt/pgi/linux86-64/7.1/mpi/lam/bin /opt/pgi/linux86-64/7.1/bin:$PATH ! for 64-bit cd examples make >& make.log cd hello ! enter one of the examples mpirun -np 4 cxxhello ! run the example
Results of the example tests can be found in the make.log above. Review this file for any errors
The PGI graphical MPI-parallel debugger, PGDBG, has not been modified to work with LAM-MPI.
None.