|
Abstract : |
High-Performance Fortran as currently defined cannot be used to solve all programming problems. However, its focus on regular problems and data-parallel algorithms stems not from a belief that it is only these problems that matter, but rather from the fact that it was in this area that it was easiest to build consensus as to what was required in a language for high-performance computing. In future work, this initial consensus will have to be extended. One direction to be considered in the next round of HPF Forum meetings is task parallelism. In this paper, we examine and illustrate the considerations that motivate the use of task parallelism. We also describe one particular approach to task parallelism in Fortran, namely the Fortran M extensions. Finally, we contrast Fortran M with other proposed approaches and discuss the implications of this work for task parallelism and high-performance languages. 1, |
