摘要
本文从开发指令级并行度ILP的角度出发,分析了超标量、超流水线处理器的体系结构特点,在此基础上给出了一个定点超标量RISC核设计。该设计采用Top-down设计方法,含三个流水执行单元,指令动态调度,实现非阻塞高速缓存non-blocking-caches机制。
By exploiting instruction-level parallelism (ILP), the characteristics of architecture of super-scalar and super-pipelined processors are analyzed. Based on the achievement of research, a design of RISC core by super-scalar and super-pipelined technique is given. According to the proposed design, three pipelined execution units are configured by top-down methodology. Instruction dynamic scheduling and non-blocking-caches mechanism are also implemented.
出处
《电路与系统学报》
CSCD
2001年第4期56-60,共5页
Journal of Circuits and Systems
基金
国家自然科学基金(69872033)
浙江省综合信息网技术重点实验室
教育部骨干教师计划资助项目
关键词
RISC核
指令
微处理器
电路设计
Instruction-level parallelism
super-scalar
super-pipelined
non-blocking-cache
reservation station.