论文部分内容阅读
根据实时信号处理应用中嵌入式计算硬件的资源受限特点,提出一种新的适合计算任务流的实时任务调度算法,以满足吞吐率为调度约束条件、以最小化流水线时延为调度目标并兼顾资源节约原则.算法为两层结构,首先采用任务插入策略得到满足吞吐率约束的中间结果;再通过基于任务复制的优化调度,分析了影响调度长度的关键任务并按照所定义的优先级构造关键边的决策树以确定需复制的关键父任务,充分利用了处理器的空闲时间片以减少时延.另外还设计了检测任务复制后出现的冗余处理器的方法.通过随机任务图仿真分析,说明该算法能以较小的处理器使用量获得满足吞吐率约束的调度结果.
According to the limited resources of embedded computing hardware in real-time signal processing applications, a new real-time task scheduling algorithm suitable for computing task flow is proposed to meet the throughput constraints of scheduling and minimize the pipeline delay as the scheduling target And take into account the principle of resource conservation.The algorithm is a two-tier structure, the task insertion strategy is used to obtain the intermediate results satisfying the throughput constraint.Furthermore, through the task-based optimization scheduling, the key tasks affecting the schedule length are analyzed and constructed according to the defined priority Key-edge decision tree to determine the key tasks to be copied, make full use of the processor’s free time slice to reduce the delay.In addition, the paper also designed a method of detecting redundant processors after task duplication.Through the random task graph simulation Analysis shows that this algorithm can get the scheduling result that satisfies the throughput constraint with a smaller processor usage.