Graspg -- An extension to Grasp2018 based on Configuration State Function Generators

TL;DR

Graspg extends Grasp2018 by incorporating configuration state function generators, significantly improving computational efficiency and memory usage for large-scale atomic structure calculations, enabling handling of extremely large CSF sets.

Contribution

The paper introduces Graspg, an extension to Grasp2018, with new CSFG-based methods, improved potential constructions, and parallelization, allowing efficient large-scale MCDHF and CI calculations.

Findings

Execution time reduced by factors of 37 to over 200.

Disk file sizes decreased by factors up to 98.

Large CSF sets (hundreds of millions) are manageable with new methods.

Abstract

The Graspg program package is an extension of Grasp2018 [Comput. Phys. Commun. 237 (2019) 184-187] based on configuration state function generators (CSFGs). The generators keep spin-angular integrations at a minimum and reduce substantially the execution time and the memory requirements for large-scale multiconfiguration Dirac-Hartree-Fock (MCDHF) and relativistic configuration interaction (CI) atomic structure calculations. The package includes the improvements reported in [Atoms 11 (2023) 12] in terms of redesigned and efficient constructions of direct- and exchange potentials, as well as Lagrange multipliers, and additional parallelization of the diagonalization procedure. Tools have been developed for predicting configuration state functions (CSFs) that are unimportant and can be discarded for large MCDHF or CI calculations based on results from smaller calculations, thus providing efficient methods for a priori condensation. The package provides a seamless interoperability with Grasp2018. From extensive test runs and benchmarking, we have demonstrated reductions in the execution time and disk file sizes with factors of 37 and 98, respectively, for MCDHF calculations based on large orbital sets compared to corresponding Grasp2018 calculations. For CI calculations, reductions of the execution time with factors over 200 have been attained. With a sensible use of the new possibilities for a priori condensation, CI calculations with nominally hundreds of millions of CSFs can be handled.