PararealF90

A lightweight special-purpose Fortran90 implementation of Parareal.

Provides implementations of Parareal in both MPI and OpenMP to allow for a comparison of the effect of different parallelisation strategies on runtime, memory footprint and energy consumption. [1]

1. Ruprecht2017_lncs

   D. Ruprecht, “Shared Memory Pipelined Parareal,” in Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 – September 1, 2017, Proceedings, F. F. Rivera, T. F. Pena, and J. C. Cabaleiro, Eds. Springer International Publishing, 2017, pp. 669–681 [Online]. Available at: https://doi.org/10.1007/978-3-319-64203-1_48

   BibTeX

   Abstract

   BibTeX entry Ruprecht2017_lncs

   ×

   @inbook{Ruprecht2017_lncs,
     author = {Ruprecht, Daniel},
     booktitle = {Euro-Par 2017: Parallel Processing: 23rd International Conference on Parallel and Distributed Computing, Santiago de Compostela, Spain, August 28 -- September 1, 2017, Proceedings},
     doi = {10.1007/978-3-319-64203-1_48},
     editor = {Rivera, Francisco F. and Pena, Tom{\'a}s F. and Cabaleiro, Jos{\'e} C.},
     pages = {669--681},
     publisher = {Springer International Publishing},
     title = {Shared Memory Pipelined Parareal},
     url = {https://doi.org/10.1007/978-3-319-64203-1_48},
     year = {2017}
   }
   

   Abstract for BibTeX entry Ruprecht2017_lncs

   ×

   For the parallel-in-time integration method Parareal, pipelining can be used to hide some of the cost of the serial correction step and improve its efficiency. The paper introduces a basic OpenMP implementation of pipelined Parareal and compares it to a standard MPI-based variant. Both versions yield almost identical runtimes, but, depending on the compiler, the OpenMP variant consumes about 7% less energy and has a significantly smaller memory footprint. However, its higher implementation complexity might make it difficult to use in legacy codes and in combination with spatial parallelisation.