Automatic Generation of Multi-Dimensionally Partitioned Parallel CFD code in a Parallelisation Tool

Evans, E. W.; Johnson, S. P.; Leggett, P. F.; Cross, M.

doi:10.1016/b978-044482849-1/50063-4

Cited by 6 publications

(2 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The complexity of the procedures depends on the complexity of the problems to be dealt with. For a simple domain partitioning problem, Evans developed a toolkit known as the Computer Aided Parallelization Tools(CAPTools) to transfer a scalar code to a form suitable for parallel implementation with message passing calls [5]. For the complex problems that have different local meshing index systems, it is difficult and time consuming to deal with the inner boundary data exchange.…”

Section: Introductionmentioning

confidence: 99%

A General Sub-Domain Boundary Mapping Procedure For Structured Grid CFD Parallel Computation

Wang

Zha

2007

25th AIAA Applied Aerodynamics Conference

View full text Add to dashboard Cite

In the present work, a general sub-domain boundary mapping procedure has been developed for arbitrary topology multi-block structured grids with grid points matched on sub-domain boundaries. The interface of two adjacent blocks is uniquely defined according to each local meshing index system (MIS) which is specified independently. A pack/unpack procedure based on the definition of the interface is developed to exchange the data in 1D array to minimize the communication amount. A safe send/receive procedure is built to remove the possibility of communication block and achieve an optimum parallel computation efficiency. The procedure is applied to parallelize an in house 3D Navier-Stokes code. The message passing interface (MPI) protocol is used for the data communication. The implementation of this newly developed parallelization procedure is straightforward. The programming work to convert a sequential code to parallel code using this procedure is minimal. The numerical experiments on an MPI based computer cluster show that the general mapping algorithm is robust and has high parallel computation efficiency.

show abstract

Section: Introductionmentioning

confidence: 99%

A General Sub-Domain Boundary Mapping Procedure For Structured Grid CFD Parallel Computation

Wang

Zha

2007

25th AIAA Applied Aerodynamics Conference

View full text Add to dashboard Cite

show abstract

“…The computer aided parallelisation tools [23,24] (CAPTools) have been developed by the Parallel Processing Research Group at the University of Greenwich to handle complex application codes (both new and dusty deck) in a structured (regular) grid context using mesh decomposition parallelisation techniques [25,26]. CAPTools exploits advanced interprocedural symbolic dependence analysis [27][28][29] within a highly sophisticated user environment [23] that allows interaction with the tools, to direct and enhance the resulting parallel code.…”

Section: Introductionmentioning

confidence: 99%