PCTDSE: A parallel Cartesian-grid-based TDSE solver for modeling laser-atom interactions

We present a parallel Cartesian-grid-based time-dependent Schrodinger equation (TDSE) solver for modeling laser-atom interactions. It can simulate the single-electron dynamics of atoms in arbitrary time-dependent vector potentials. We use a split-operator method combined with fast Fourier transforms (FFT), on a three-dimensional (3D) Cartesian grid. Parallelization is realized using a 2D decomposition strategy based on the Message Passing Interface (MPI) library, which results in a good parallel scaling on modern supercomputers. We give simple applications for the hydrogen atom using the benchmark problems coming from the references and obtain repeatable results. The extensions to other laser-atom systems are straightforward with minimal modifications of the source code. Program summary Program title: PCTDSE Catalogue identifier: AFBM_v1_0 Program summary URL: http://cpc.cs.qub.ac.ukisummaries/AFBM_v1_0.html Program obtainable from: CPC Program Library, Queen's University, Belfast, N. Ireland Licensing provisions: BSD 3- Clause No. of lines in distributed program, including test data, etc.: 398662 No. of bytes in distributed program, including test data, etc.: 6355779 Distribution format: tar.gz Programming language: Fortran 2003. Computer: Distributed memory machines. Operating system: Unix-like system. RAM: Depends on the size of the Cartesian grid Classification: 2.1, 2.2, 2.5. External routines: 2DECOMP82FFT; FFTW; MPI Nature of problem: Simulate the single-electron dynamics of atoms under the strong fields of modern lasers according to the time-dependent Schrodinger equation. Solution method: The package solves the TDSE in the FFT-split-operator method, employing the split-operator method to approximate the time evolution operator and fast Fourier transforms to calculate the spatial derivatives. Restrictions: The code is restricted to problems where the atoms are in the single-active-electron approximation and the lasers are in the dipole approximation. It is also limited by the CPU time and memory that one can afford. Unusual features: We adopt the "parallel strategy where the Cartesian grid is distributed among processors using a 2D decomposition, which has no limitation for large-scale simulations. Running time: The running time depends on the size of the grid, the number of time step, the number of processors, and the choice of the processor grid, ranging from a few hours to several days. (C) 2016 Elsevier B.V. All rights reserved.