Efficient support for irregular applications on distributed-memory machines
PPOPP '95 Proceedings of the fifth ACM SIGPLAN symposium on Principles and practice of parallel programming
Dynamic Partitioning of Non-Uniform Structured Workloads with Spacefilling Curves
IEEE Transactions on Parallel and Distributed Systems
Proceedings of the eighth annual ACM symposium on Parallel algorithms and architectures
Efficient run-time support for irregular block-structured applications
Journal of Parallel and Distributed Computing - Special issue on irregular problems in supercomputing applications
Parallelization of Irregular Problems Based on Hierarchical Domain Representation
HPCN Europe 2000 Proceedings of the 8th International Conference on High-Performance Computing and Networking
A Hierarchical Approach to Irregular Problems (Research Note)
Euro-Par '00 Proceedings from the 6th International Euro-Par Conference on Parallel Processing
Adaptive Multigrid Methods in MPI
Proceedings of the 7th European PVM/MPI Users' Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface
Global trees: a framework for linked data structures on distributed memory parallel systems
Proceedings of the 2008 ACM/IEEE conference on Supercomputing
Parallel hierarchical radiosity: the PIT approach
PARA'04 Proceedings of the 7th international conference on Applied Parallel Computing: state of the Art in Scientific Computing
Hi-index | 0.00 |
A problem is irregular if its solution requires the computation of some properties for each of a set of elements irregularly distributed in a domain of interest. These problems satisfy a locality property because the properties of an element depend upon those of a few other elements, its neighbors, according to a dynamic, problem dependent stencil. The development of parallel algorithms for irregular problems on distributed memory architectures is not trivial, because the irregularity and the dinamicity of the distribution of the elements in the domain require complex strategies to manage the mapping of elements onto the processing nodes and to implement the processing nodes cooperation. This paper introduces PIT, a library to simplify the parallelization of irregular problems. The key assumption underlying the definition of PIT is that both the sequential and the parallel version of the application are structured in terms of operations on a tree that describes the distribution of the elements in the domain. In the parallel version, the tree is handled in parallel through the functions supplied by PIT in a way that is transparent to the user and that preserves most of the sequential code.