自调整附加权动态仲裁算法

An Adaptive and Weighted Dynamic Arbitration Algorithm

基于总线的SoC存在多模块对总线的抢占问题,仲裁机制可解决抢占引起的冲突和竞争,提高系统性能,为此提出一种自调整附加权动态仲裁算法.首先在每轮仲裁开始时预先判断工作环境,针对不同工作环境改进请求信号,判定工作重点是快速响应或复杂目的带宽分配;然后结合RR仲裁算法和FP仲裁算法的优势完成仲裁.在NINP模型下,采用65 nm CMOS工艺的Xilinx Virtex5进行验证的结果表明,与传统的仲裁算法相比,文中算法具有更好的带宽分配功能,避免了“饥饿”和“独占”现象,总线利用率提高了11.3％～56.3％;该算法逻辑简单,容易实现,能满足多种环境下基于总线的SoC应用.

Arbitration mechanism can deal with contentions caused by the bus preemption of multiple masters in SoC, thus the performance of the system is improved. An adaptive and weighted dynamic arbitration algorithm is proposed. At the beginning of each arbitration cycle, the proposed arbitration algorithm cheeks working environment, then requests can be changed correspondingly to achieve quick response or complicated bandwidth allocation. Arbitration can be accomplished when combing the advantages of RR and FP arbitration algorithms. In NINP model, verifications are made on Xilinx Virtex5 with 65nm CMOS technology. The experimental results show that the proposed arbitration algorithm is better in bandwidth allocation and can avoid ＂starvation＂ and ＂monopolization＂ of buses in comparison with conventional arbitration algorithms. Furthermore, the bus utilization increases by 11. 35-56.3%. The proposed arbitration algorithm is simple in logic, easy to implement and suitable in various environments of SoC applications based on buses.