Parallel Haskell implementations of the N‐body problem

Abstract

SUMMARYThis paper provides an assessment of high-level parallel programming models for multi-core programming by implementing two versions of the n-body problem. We compare three different parallel programming models based on parallel Haskell, differing in the ways how potential parallelism is identified and managed. We assess the performance of each implementation, discuss the sequential and parallel tuning steps leading to the final versions, and draw general conclusions on the suitability of high-level parallel programming models for multi-core programming. We achieve speed-ups of up to 7.2 for the all-pairs algorithm and up to 6.5 for the Barnes-Hut algorithm on an 8-core machine.

Lazy data-oriented evaluation strategies

“…We adapt the list-based algorithm used in [27] by performing the main map operation in the second phase over a quad-tree data structure. This enables us to use our more advanced strategies defined here, while comparing the performance with an already well-tuned implementation.…”

A framework for argument-based task synchronization with automatic detection of dependencies

“…Nevertheless, the exhaustive exploitation of all this implicit parallelism would involve much overhead. For this reason, in practice functional languages provide mechanisms to explictly label those computations whose parallel evaluation can be advantageous [23]. While some approaches [24,25,26] lead to explicit communications and synchronizations, in others the user just identifies the parallel tasks, letting the compiler or runtime take care of the low level details.…”

SICSA multicore challenge editorial preface

“…Further information is available at the website † † . [31] N -body x86 Single-assignment C Goli et al [32] N -body x86 and NVidia GPU Skeletons (FastFlow, SkePU and Thrust) Totoo et al [33] N -body x86 Parallel Haskell (GpH/Eden)…”