摘要
本文讨论了源串行LISP语言的并行性开发,包括任务的划分和表示、并行处理策略、任务的环境控制、任务的调度和通讯,并给出了提高任务并行度的特殊函数处理、伪结果控制、自定义函数任务提升策略等算法,上述思想已在实验系统MTASKLISP上初步实现.
This paper discusses the exploitation of parallelism in sequential LISP source programs. The problems studied includes partitioning and representation of tasks, parallel processing strategy, task environment control, scheduling and communication. Algorithms concerning the improvement of parallelism degree, such as special function processing, pseudo result control, and promotion strategy of self-defined function tasks, are also given. All these ideas are implemented in an experimental system named MTASKLISP.
出处
《计算机学报》
EI
CSCD
北大核心
1992年第7期546-551,共6页
Chinese Journal of Computers
基金
航空航天部基金
关键词
LISP语言
并行处理
Artificial intelligence, parallel process, LISP language, task tree.