Part of: Refactoring a climate simulation miniapp to use a PGAS programming model (2 objects) Next
Refactoring a climate simulation miniapp to use a PGAS programming model [video]
Refactoring a climate simulation miniapp to use a PGAS programming model [presentation]
Refactoring a climate simulation miniapp to use a PGAS programming model [video]
Times viewed: 131
Description
This presentation was delivered by a student in the Summer Internships in Parallel Computational Science Program (SIParCS) at the National Center for Atmospheric Research (NCAR). SIParCS embeds graduate and undergraduate students as summer interns in NCAR's Computational and Information System Laboratory, offering them significant hands-on R&D opportunities in high performance computing (HPC) and related fields that use HPC for scientific discovery and modeling. Partitioned global address space (PGAS) programming models are designed to improve programmer productivity when implementing parallel software. Despite this aim, however, PGAS models have not currently gained wide-scale adoption. This is due in part to the lack of studies that examine the applicability of these programming models to large-scale codes. In this work we use the Co-Array Fortran programming model to implement the border-exchange algorithm within a climate simulation miniapp. Miniapps are small applications that include complicating details seen in larger-scale codes.