A Skeleton Library

Abstract

Today, parallel programming is dominated by message passing libraries such as MPI. Algorithmic skeletons intend to simplify parallel programming by increasing the expressive power. The idea is to offer typical parallel programming patterns as polymorphic higher-order functions which are efficiently implemented in parallel. The approach presented here integrates the main features of existing skeleton systems. Moreover, it does not come along with a new programming language or language extension, which parallel programmers may hesitate to learn, but it is offered in form of a library, which can easily be used by e.g. C and C++ programmers. A major technical difficulty is to simulate the main requirements for a skeleton implementation, namely higher-order functions, partial applications, and polymorphism as efficiently as possible in an imperative programming language. Experimental results based on a draft implementation of the suggested skeleton library show that this can be achieved without a significant performance penalty.1 We assume some familiarity with MPI and C++.

An Implementation of a Parallel Generalized Branch and Bound Template

“…It was proposed by M. Cole in his PhD thesis [13], see also [6,15,12]. Any parallel algorithm template must fully or partially specify the main features of a parallel algorithm: partitioning, communication, agglomeration and mapping.…”

Automatic Task-Based Code Generation for High Performance Domain Specific Embedded Language

“…Even if a large number of skeletons has been proposed in the litterature [24,12], NT 2 focuses on three data-oriented skeletons:…”

Streaming Applications

“…There are several parallel programming frameworks based on design patterns or skeletons, among which CO 2 P 3 S [86], eSkel [41], Lithium [6], P 3 L [17], and the Kuchen Skeleton library [78]. Since these frameworks focus on parallel programming in general, they have minimal support for streaming applications.…”