Parallelization of MIN3P-THCm: A high performance computational framework for subsurface flow and reactive transport simulation. (September 2017)

• Record Type: Journal Article
• Authors: Su, Danyang; Ulrich Mayer, K.; MacQuarrie, Kerry T.B.
• Is Part Of: Environmental modelling & software. Volume 95(2017)
• Page Start: 271
• Page End: 289
• Publication Date: 2017-09
• Subjects: MIN3P-THCm -- ParMIN3P-THCm -- Reactive transport -- Parallel computing -- Domain decomposition -- PETSc; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2017.06.008
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: This paper presents the development of ParMIN3P-THCm, a parallel version of the reactive transport code MIN3P-THCm, which can run efficiently on machines ranging from desktop PCs to supercomputers. Parallelization of ParMIN3P-THCm was achieved through the domain decomposition method based on the PETSc library. The code has been developed from the ground up for parallel scalability and has been tested for up to 768 processors with problem sizes up to 100 million unknowns, showing strong scalability in modeling large-scale reactive transport problems. The total speedup tends to be ideal and near linear up to 768 processors when the degrees of freedom per processor is larger than 8000–15, 000, depending on the relative complexity of the reactive transport and flow problems. The improved code efficiency allows refining of the model discretization in both space and time and will facilitate 3D simulations that were impractical to carry out with the sequential version of MIN3P-THCm. Highlights: An efficient parallel version of MIN3P-THCm for reactive transport simulation has been developed. Global implicit reactive transport solution based on parallel domain decomposition method. Strong scalability on machines ranging from desktop PCs to distributed-memory supercomputers was achieved. The code was developed based on the PETSc framework with multiple parallel solvers available.