Algorithmic Skeletons - The Muenster Skeleton Library (Muesli) (Muesli)

Basic data for this project

Type of projectOwn resources project
Duration at the University of Münster01/02/2002 - 31/01/2025

Description

The Muenster Skeleton Library Muesli is a C++ template library enabling the hassle-free programming of multi-node, multi-core cluster computers by implementing the concept of so-called algorithmic skeletons (skeletons for short). In essence, skeletons are higher order functions and encapsulate typical parallel computation/communication patterns. By predefining them in a library, the user does not need to bother about programming problems typically encountered when using MPI and/or OpenMP manually, such as deadlocks, starvation, mutual exclusion etc. Instead, when using Muesli, all communication details are encapsulated inside our library, such that parallel programming is taken to a higher level of abstraction. Users do not need to bother with MPI and/or OpenMP, but can simply implement parallel programs as if they were sequential thanks to the SPMD execution model underlying our library. In essence, Muesli makes parallel programming easier, safer, and less error-prone.

KeywordsHigh Level Parallel Programming; Algorithmic Skeletons
Website of the projecthttp://muesli.uni-muenster.de/

Project management at the University of Münster

Kuchen, Herbert
Practical Computer Science Group (PI)

Research associates from the University of Münster

Herrmann, Nina
Practical Computer Science Group (PI)
Menezes, Breno
Practical Computer Science Group (PI)