Advanced parallel programming models research and development opportunities.

Abstract

There is currently a large research and development effort within the highperformance computing community on advanced parallel programming models. This research can potentially have an impact on parallel applications, system software, and computing architectures in the next several years. Given Sandia's expertise and unique perspective in these areas, particularly on very large-scale systems, there are many areas in which Sandia can contribute to this effort. This technical report provides a survey of past and present parallel programming model research projects and provides a detailed description of the Partitioned Global Address Space (PGAS) programming model. The PGAS model may offer several improvements over the traditional distributed memory message passing model, which is the dominant model currently being used at Sandia. This technical report discusses these potential benefits and outlines specific areas where Sandia's expertise could contribute to current research activities. In particular, we describe several projects in the areas of high-performance networking, operating systems and parallel runtime systems, compilers, application development, and performance evaluation. 3 AcknowledgmentWe would like to express our appreciation to the people who helped us during the preparation of this technical report. Sue Goudy reviewed early drafts and identified several important Sandia applications. Thomas Christopher provided background on earlier parallel programming models and suggested the idea of a possible implementation of UPC on the PIM architecture. Mike Heroux shared his invaluable experiences about suitable applications for UPC and directed us to important applications, such as the Trilinos project (the Epetra package) and Co-array Fortran applications, that may be suitable for this work. Neil Pundit reviewed several drafts and provided input and guidance. Jonathan L. Brown assisted with editing and revising this report, provided several UPC coding examples, and conveyed early experiences with the language. Last but not least, Bill Camp provided many insightful comments and constructive suggestions, which helped to give a more complete and practical perspective to this report. 4