Process Partitioning and Scheduling on Shared Memory Systems

• Main Authors: Tsai-Yih Kuo, 郭蔡義
• Other Authors: Jong-Jiann Shieh
• Format: Others
• Language: en_US
• Published: 1994
• Online Access: http://ndltd.ncl.edu.tw/handle/34836120849524761268

碩士 === 大同工學院 === 資訊工程研究所 === 82 === A scheduling strategy for shared memory multiprocessor systems is described. The objective of our research is to reduce the total execution time of the application programs that are to be executed on the target shared memory multiprocessor systems. Generally, the execution time includes the program codes computing time and the data communication time in a multiprocessor system. The time of data communication by the common memory would be decreased by collecting instructions with data dependence together. A program should be translated into many blocks, called parallel execution blocks, and blocks can be executed concurrently to reduce total program computing time. An application program is compiled to the intermediate code and represented by a Control Dependence Graph (CDG). The partition procedure translates CDG into many parallel execution program blocks, and then the scheduler process these blocks. The scheduler takes the blocks computing time and the interblocks communication time into account to allocate these blocks on target systems. The partitioner and the scheduler have been implemented. Some application programs have been tested.