摘要
在行为级综合中,一般在调度与分配之前先将系统的描述转换成一个数据流程图 ( DFG)。研究中发现在资源约束条件下有最小操作数的固定的 DFG并不一定获得一 个最短时间的调度。我们研究了一个新的调度算法──变换数据流程图法,它改变图的 拓朴结构,并且同时进行调度。实例证明我们的算法所获得的调度优于原始的DFG。
In behavior synthesis, the behavior of the system is always transformed to a DFG before scheduling and allocation. In our research, it is found that under resource constraint, the fixed DFG with the minimal operation does not necessarily gain the best scheduling. Here, we present a new scheduling algorithm-Transformed DFG. It changes the topology of the DFG in the process of scheduling. The examples show our algorithm obtain a hotter scheduling than that of the primary DFG.
出处
《计算机辅助工程》
2000年第1期70-74,共5页
Computer Aided Engineering