Mapping fortran programs to single assignment semantics for efficient parallelization

Abstract

This paper presents Mustang, a system that automatically parallellizes Fortran programs by mapping them to single assignment semantics. Specifically, sequential Fortran source programs are translated into IF1, a machine-independent dataflow graph description language that is the intermediate form for the SISAL language. During this translation, Parafrase 2 is used to parse the source program perform dependency analysis and to detect opportunities for parallelization which are then explicitly introduced into the IF1 program. The resulting IF1 program is then processed by the Optimizing SISAL Compiler which produces parallel executables on multiple target platforms. A working prototype has been developed and tested. The execution results of several Livermore Loops are presented and compared against Fortran and SISAL implementations on two different platforms. The initial results obtained provide proof of concept that Fortran can be mapped to Single Assignment Semantics without sacrificing efficiency. © World Scientific Publishing Company.