指令级并行的数学模型

The Mathematical Model of Instruction Level Parallelism

目前，提高CPU性能的一种重要方法是发掘指令级并行性。在CPU里设置多个功能部件，采用复杂的指令执行机制，使尽可能多的指令同时执行。本文用数学模型分析指令并行的一般原理，阐述指令相关的数学本质，并分析消除数据相关和控制相关的基本方法。

At present,an important method to improve the CPU performance is to explore instruction-level parallelism.Add multiple component to the processor and adopt complicated executive mechanism,so that the processor can execute instructions as many as possible at the same time.We＇ve developed a model to analyze the principle of summarize general method to eliminate data and control dependency.