A Nonannotative Approach to Distributed Data-Parallel Computing
暂无分享,去创建一个
An approach to data-parallel computing is presented which avoids annotation by introducing a type system with symmetric subtyping. The properties that are usually specified in annotations in a machine-dependent way become deducible from type signatures of data objects. The chief advantage of the method is that it caters for portability by presenting a data description in terms of algorithmic properties (most importantly symmetry of data and of access to it) rather than any machine-specific terms.
[1] Murray Cole,et al. Algorithmic Skeletons: Structured Management of Parallel Computation , 1989 .
[2] Guy E. Blelloch,et al. NESL: A Nested Data-Parallel Language (Version 2.6) , 1993 .
[3] Alexander V. Shafarenko,et al. F-Code and its Implementation: A Portable Software Platform for Data Parallelism , 1993, Comput. J..
[4] Raymond P. Polivka,et al. APL2: at a glance , 1988 .
[5] Alexander V. Shafarenko,et al. Data parallel computing - the language dimension , 1996 .