Parallelization of plasma simulation codes: gridless finite size particle versus particle in cell approach

The main features of gridless finite-size-particle codes are examined and c ompared with those of particle-in-cell ones, from the point of view of the performances that can be obtained, with respect to both the spatial-resolut ion level and the efficiency of parallel particle simulations. The particle -decomposition parallelization strategy is discussed, and its implementatio n in the framework of the High Performance Fortran paradigm is presented. I t is shown that such codes are particularly suited for particle-decompositi on parallelization on distributed-memory architectures, as they present a s trong reduction, in comparison with particle-in-cell codes, of the memory a nd computational offsets related to storing and updating the replicated flu ctuating-field arrays. (C)2000 Elsevier Science B.V. All rights reserved.