A communication library between multiple sets of MPI processes for a MPMD model

Shimosaka, Takenori; Murai, Hitoshi; Sato, Mitsuhisa

doi:10.1145/2488551.2488597

Cited by 1 publication

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We measured the basic performance of MPMPI_Send/Recv in the previous study [2]. This section shows the improved basic performance of MPMPI_Send/Recv and the experimental result of the block-based Gauss-Jordan program (BGJ) and a coupling program using multiple Laplace programs, called the Laplace Coupling Program (LCP).…”

Section: Methodsmentioning

confidence: 99%

“…As shown, Unpack is parallelized with the OpenMP loop directive. In the previous study [2], we did not use the OpenMP directives. Here, nbuf_stride is nbuf_stride1 or nbuf_stride2, and i_init_loc shows the location of the y-axis direction.…”

Section: E Pack/unpackmentioning

confidence: 99%

“…We measured the performance of the same simple communication benchmark program used in the previous study [2]. The benchmark program comprised two MPMPI functions that send/receive two-dimensional distributed arrays on different node arrays.…”

Section: A Preliminary Performance Of Mpmpimentioning

confidence: 99%

See 2 more Smart Citations

A Design of a Communication Library between Multiple Sets of MPI Processes for MPMD

Shimosaka

Murai

Sato

2014

2014 IEEE 17th International Conference on Computational Science and Engineering

Self Cite

View full text Add to dashboard Cite

An MPMD programming model is widely used as a master-worker program or a coupling program for multiple physical models. To utilize recent high-end parallel computers having more than several thousand nodes, we propose the communication library MPMPI between different multiple sets of MPI processes in the MPMD model. In particular, we present MPMPI interfaces that include interfaces for a PGAS language and the basic performance of MPMPI functions. As benchmark programs of the MPMPI library, we evaluated the performance of a master-worker program and a weak coupling program. As a result, we found that Pack/Unpack has a large influence on the performance of MPMPI functions, the interface of MPMPI functions can easily be used in these benchmark programs written in XcalableMP PGAS language, and the performances of the master-worker and weak coupling benchmark programs using the basic MPMPI functions are practical under the conditions of this paper.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: E Pack/unpackmentioning

confidence: 99%