The implementation of MPI-2 one-sided communication for the NEC SX-5
Proceedings of the 2000 ACM/IEEE conference on Supercomputing
Optimizing noncontiguous accesses in MPI – IO
Parallel Computing
Flattening on the Fly: Efficient Handling of MPI Derived Datatypes
Proceedings of the 6th European PVM/MPI Users' Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface
Fast Parallel Non-Contiguous File Access
Proceedings of the 2003 ACM/IEEE conference on Supercomputing
Constructing MPI Input-output Datatypes for Efficient Transpacking
Proceedings of the 15th European PVM/MPI Users' Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface
Mapping functions and data redistribution for parallel files
The Journal of Supercomputing
Automatic memory optimizations for improving MPI derived datatype performance
EuroPVM/MPI'06 Proceedings of the 13th European PVM/MPI User's Group conference on Recent advances in parallel virtual machine and message passing interface
Benchmarking one-sided communication with SKaMPI 5
PVM/MPI'05 Proceedings of the 12th European PVM/MPI users' group conference on Recent Advances in Parallel Virtual Machine and Message Passing Interface
Hi-index | 0.01 |
Based on a construction of socalled input-output datatypes that define a mapping between non-consecutive input and output buffers, we outline an efficient method for copying of structured data. We term this operation transpacking , and show how transpacking can be applied for the MPI implementation of one-sided communication and MPI-IO. For one-sided communication via shared-memory, we demonstrate the expected performance improvements by up to a factor of two. For individual MPI-IO, the time to read or write from file dominates the overall time, but even here efficient transpacking can in some scenarios reduce file I/O time considerably. The reported results have been achieved on a single NEC SX-8 vector node.