Closed streeve closed 4 years ago
This PR fixes LAMMPS read_data input for MPI runs, where data was previously read incorrectly causing segfaults.
This also avoids allocation N_global atoms per rank (resizing if the guess for N_local is too small)
This is related to #28, which only partially fixed the issue.
This PR fixes LAMMPS read_data input for MPI runs, where data was previously read incorrectly causing segfaults.
This also avoids allocation N_global atoms per rank (resizing if the guess for N_local is too small)
This is related to #28, which only partially fixed the issue.