International Journal of High Performance Computing and Networking · 2004 · 66 citations · 17 references
ComputingComputer ScienceHigh-performance ComputingCompilersParallel ComputingProgramming LanguagesInterprocess CommunicationParallel Programming ModelsShared-memory MultiprocessingAssembly Language ProgrammingCompiler SupportGlobal Memory Access PatternsHigh PerformanceHigh-performance Computing SystemsMemory Access PatternsOne-sided CommunicationParallel Programming
MPI support is nearly ubiquitous on high-performance systems today and is generally highly tuned for performance. It would thus seem to offer a convenient 'portable network assembly language' to developers of parallel programming languages who wish to target different network architectures. Unfortunately, neither the traditional MPI 1.1 API nor the newer MPI 2.0 extensions for one-sided communication provide an adequate compilation target for global address space languages, and this is likely to be the case for many other parallel languages as well. Simulating one-sided communication under the MPI 1.1 API is too expensive, while the MPI 2.0 one-sided API imposes a number of significant restrictions on memory access patterns that would need to be incorporated at the language level, as a compiler cannot effectively hide them given current conflict and alias detection algorithms.
17
MPI: A Message-Passing Interface Standard
Message P Forum · 1994 · 2.7K citations
Co-array Fortran for parallel programming
Robert W. Numrich, John Reid · ACM SIGPLAN Fortran Forum · 1998 · 671 citations
Titanium: a high-performance Java dialect
Kathy Yelick, Luigi Semenzato, Geoff Pike et al. · Concurrency Practice and Experience · 1998 · 456 citations
A performance analysis of the Berkeley UPC compiler
Wei-Yu Chen, Dan Bonachea, Jason Duell et al. · 2003 · 127 citations · Full text