Abstract
In order to achieve high productivity in parallel programming, it becomes possible for programmers to generate parallelized programs with inter-process communication by parallelizing compilers such as HPF and PGAS compilers. However, it is required for programmers to specify distribution of arrays by inserting directives in the programs in high level parallel programming languages. Therefore, this paper proposes a new method to determine array data distribution automatically. One of the conventional data distribution methods is the Component Afflinity Graph (CAG) based method where relations between multiple array dimensions referenced in nested loops are analyzed. With our method, it is possible to determine more appropriate data distribution of arrays in multiply nested loops by analyzing relations of dimensions of the arrays while considering parallelisim of the loops. With the proposed method, sequential Fortran programs are automatically translated into HPF programs with data distribution directives. The genarated HPF program was parallelized by the PGI HPF compiler and executed on a PC cluster. We confirmed that scalability of the benchmark is close to those of NPB3.0-HPF.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
