Execution of jx and MPI parallelization

To execute the program 'jx', it is necessary to type the command in the following orders.

     % ./jx fe2.scfout jx.config > jx.log

The first argument should be a scfout file generated by OpenMX. The second argument is the configuration file explained above.

Also, 'jx' supports the MPI parallelization for periodic calculations.

     % mpirun -np 2 ./jx fe2.scfout jx.config > jx.log

Note that MPI parallelization of `jx' is supported only for the eigenvalue solver `band' and not for `cluster'. We also have to note that the parallelization is done only for the $k$ vector, and when the parallelization number becomes larger than the total number of $k$-points it would result in unnecessary consumption of computational resources.