Tiling multidimensional iteration spaces for nonshared memory machines
Proceedings of the 1991 ACM/IEEE conference on Supercomputing
Optimal orthogonal tiling of 2-D iterations
Journal of Parallel and Distributed Computing
Skeletons for Data Parallelism in p3l
Euro-Par '97 Proceedings of the Third International Euro-Par Conference on Parallel Processing
Euro-Par '98 Proceedings of the 4th International Euro-Par Conference on Parallel Processing
A Skeleton for Parallel Dynamic Programming
Euro-Par '99 Proceedings of the 5th International Euro-Par Conference on Parallel Processing
Exploiting Advanced Task Parallelism in High Performance Fortran via a Task Library
Euro-Par '99 Proceedings of the 5th International Euro-Par Conference on Parallel Processing
| Hi-index | 0.00 |
The portability of parallel programs has involved lot of effort during the last decade. PVM and MPI have greatly contributed to solve this drawback and nowadays most parallel programs are portable. However, the portability of the efficiency suffers, in many cases, from inherent effects of the target architectures. The optimal mapping of a parallel program is strongly dependent on the granularity and network architecture. We broach the problem of finding the optimal mapping of pipeline MPI programs. We propose an analytical model that allows an easy estimation of the parameters needed to obtain the mapping. The model is capable to be introduced into tools to produce this mapping automatically. Both the accuracy of the model and the optimal efficiency of the algorithm found are contrasted on a pipeline algorithm for the Path Planning Problem.