为了提高"萤火虫2号GPU"中的指令和数据的传输速度,设计了一款连接段操作FOP模块、2D图像加速器与帧缓冲区之间的像素cache.它采用了最近最少使用(LRU)的替换算法和二路组相连的映射方式,成功地接收来自段操作和2D图像加速器发送的指令、数据以及地址,并且作为高速缓存将数据传递给存储管理器(MMU).
In order to improve the speed of instructions and data transfer in "firefly 2nd GPU",we designed a pixel cache connect the FOP,2Dgraphics accelerator and frame buffers.It uses the LRU algorithm and two ways group connected mapping mode,successfully received from FOP and 2Dgraphics accelerator sends instructions,datas,and address,as well as send them to Storage Manager(MMU).