基于Microblaze处理器的浮点内积运算设计  

Design of floating-point inner product based Microblaze processor

作　　者：施浩[1] 姚放吾[1] 

机构地区：[1]南京邮电大学计算机学院,江苏南京210003

出　　处：《微计算机信息》2011年第5期206-207,215,共3页Control & Automation

摘　　要：浮点内积运算在信号处理与图像处理中有着广泛的应用,本文利用软核处理器灵活性和可扩展性的特点,介绍了基于Microblaze处理器的浮点内积运算结构,设计采用IEEE-754双精度浮点数,通过对DSA电路改进设计出了适合于内积运算的累加电路结构。通过EDK设计平台,在SOPC系统中把内积运算单元通过FSL总线挂载到Microblaze软核处理器上,实现了硬件单元的调用。floating-point inner product is widely applied in signal processing and image processing,this paper introduces architecture for floating-point inner product,this design is fully used of the flexible and scalability in software processor,this paper also introduces one reduction circuit which applies in floating-point inner product through improving DSA circuit.At last,we connect this computing unit to processor through FSL bus,and realize the soft-core processor called the hardware unit.

关 键 词：MICROBLAZE处理器 浮点内积运算 FSL总线 

分 类 号：TP332.2[自动化与计算机技术—计算机系统结构]