摘要
文章研究了存储控制器中的访存调度策略,提出了基于优先级的访存调度算法。首先使用遗传算法建立有效的数据源,然后对得到数据源应用统计进行调度优先级挖掘,共获取三个优先级别,这样仅使用这三个优先级构造调度算法进行访存序列调度。实验结果表明,提出的算法很好地降低了访存序列的运行时间,优化效果接近于文献[4]中提出的贪婪访存调度算法,但算法运行时间却远小于后者。
In this paper, the memory scheduling strategy of the memory controller is mainly studied, and propose a new scheduling method which is based on priority. Firstly, Genetic Algorithm (GA) is used to generate effective data resource, and then carry through priority mining by the statistics of data resource. At last, three priorities are got. So this paper uses these priorities to implement scheduling algorithm. The experiments show that our algorithm is quite efficient in total time of memory access sequence. The result is quite close to Greedy Algorithm which was used in the reference [4]. But the executing time of our algorithm is much less than Greedy Algorithm.
出处
《计算机与数字工程》
2009年第8期12-16,共5页
Computer & Digital Engineering
基金
国家"八六三"项目(编号:2009AA01Z428)资助
关键词
遗传算法
数据源
访存调度
调度优先级
genetic algorithm, data resource, memory scheduling, scheduling priority