DOWNLOAD
Elmo-Comp software is available at ElMo-Comp download
Elmo-Knock software is available at ElMo-Knock download
COMPILATION AND INSTALLATION
Software made available at this website is implemented in C++ using standard libraries available in a Linux/UNIX system. It may as well be run under CygWin in Windows OS.Serial program
To compile and install the serial version of software, run the configure script as:
$ ./configure or $ ./configure CXX=g++ if you want to specify the C++ compiler name explicitly.
Further, after configuration, compile the files into executables 'elmocomp' and 'elmoproc':
$ make
To install the executables into the standard directory run:
$ make install
To unstall it run:
$ make uninstall
Default size of the word size is set to 64-bits. It may somewhat save memory consumption if the compilation is requested for 32-bit word size using --enable-32bit as:
$./configure --enable-32bit
Parallel program (MPI only)
For parallel software version make sure that you have MPI installed and that path to library and header files are set up (e.g. PATH, LD_LIBRARY_PATH). Elementary mode sotware was tested using OpenMPI, Platform MPI and Intel MPI. Different MPI wrapper may be used with different library versions (e.g. mpiCC, mpicxx)
If the MPI compiler wrapper for C++ is mpiCC then the configure script should be called as:
$ ./configure --with-mpi=yes CXX=mpiCC
Option --with-mpi=yes is required, as default is that the program is compiled for serial execution. You cannot configure the parallel software without passing --with-mpi=yes option.
Following the successfull configuration, the compilation should be invoked as:
$ make
Depending on the parallel environment where the software is running, one should either request interactive execution from command line or create the job script (such as PBS) and submit for execution.
Parallel program (MPI + Global Arrays)
Installation instructions for this version are given at ElMo-Comp (GA) . It is recommended to use GA-enabled software for larger networks as it may have better utilization of memory.
Last change: August 2012