We propose an abstraction, named BEC, to enable Global Address Space (GAS) capabilities for parallel programming in SPMD style . It is a portable lightweight approach for incremental acceptance of the GAS model, along an evolution path that leverages existing infrastructures and maintains backward compatibility with existing programming methods and environments . It assists migration of legacy applications thereby encouraging their expert programmers to adopt the new model.In addition, it provides for some of the unaddressed needs, such as efficient support for high-volume fine-grained and random communications, which are common in parallel graph algorithms, sparse matrix operations, and large scale simulations . The idea behind BEC is that messages are aggregated by a runtime library for bulk transport to handle such unpredictable communication patterns . Data from initial experiments with a prototype communication bundling library using the Bundle-Exchange-Compute (thus motivating the name BEC) programming style shows that this approach scales well. As examples of suitable BEC applications, we present sparse matrix kernels for multiplication and overlapping Schwarz preconditioning [5,11] . We also discuss solid mechanics material contact [1,18] with abundant irregular, fine-grained communication. BEC can be used as an enhancement to existing environments such as MPI . It can also function as an intermediate language [14] to other high level GAS languages such as PRAM C [8] and UPC [30] . Furthermore, it can serve as a bridge between programming models such as virtual shared memory and message passing. pointed out the link between the BSP model and BEC . Thanks to Ron Brightwell and Rolf Riesen for helpful discussions on OS, MPI, and network issues . Mike Glass, Kevin Brown, and Courtenay Vaughan helped us a great deal by providing us with valuable insight into internal workings of Sandia software for simulation of material contacts. Bruce Hendrickson and Steve Plimpton taught us about algorithms used in Sandia's contact libraries . Doug Doerfler and Brice Fisher asked pertinent questions which motivated us to formalize the BEC model . This research used resources

Scientific Apps & User Support (4326) Stefan Domino Thermal/Fluid Computational Engineering Sciences (1541) Amalia Black V&UQ Processes (1544) Anand Ganti Advanced Networking Integration (4336) BOLD: Core 'Performance Modeling and Analysis Team' (PMAT) members

A parallel programming model, BEC, was proposed in [1], to enable Global Address Space (GAS) capabilities for programming in SPMD style . It is a portable light-weight approach for incremental acceptance of the GAS model, along an evolution path that leverages existing infrastructures . It assists migration of legacy applications thereby encouraging their expert programmers to adopt the new model . BEC also provides for some unaddressed needs, such as efficient support for high-volume fine-grained random communications . BEC can be used as an enhancement to existing environments such as MPI.This report presents a scheme for a compiler to translate high level GAS languages PRAM C ([2]) and UPC ([9]) into BEC . Since PRAM C implements the theoretical Parallel Random Access Machine (PRAM) model [5] and BEC implements the Bulk Synchronous Parallel (BSP) model [10], such a translation by a compiler is theoretically significant, because it is the first time that a program in PRAM semantics (fined-grained parallelism) is translated into program in BSP style (coarse-grained parallelism) . This provides a bridge from the PRAM model (considered impractical) to the BSP model (considered practical) . Because BEC leverages infrastructures on existing platforms (such as *email : spgoudy@sandia.gov temail : zwen@sandia .gov 3 MPI or SHMEM [7,6]), this translation scheme enables a new and alternative approach for higher level GAS language implementations which can avoid heavy investment in re-inventing much of the same communication capabilities.4 Acknowledgment