面向IXP网络处理器的内联优化

内联优化是一种有效的编译优化技术，它通过将函数体直接嵌入到调用点来消除函数调用开销。然而，网络处理器特殊的体系结构对内联优化提出了新的要求，需要新的技术辅助传统内联优化来更好地适应这种特殊的体系结构。本文描述了如何利用关键路径提取技术和迭代编译技术对传统内联优化技术进行扩充和改造，来更好地适应IXP体系结构。实验数据表明，改进后的内联优化能够有效地提高网络系统的性能。

Function Inlining for IXP Network Processor

Function inlining has been proposed as an efficient compiler optimization technique.It expands a function call site into the actual implementation of the function.This reduces overhead associated with the function call.However, current inlining techniques do not meet the demands of network processors,e.g.Intel's IXP series.The novel archi tecture features of IXP call for new techniques for adapting the traditional function inlining.This paper presents how to extend function inlining technique for IXP architecture using critical path extraction and iterative compilation.The ex- perimental result shows that the extended function inlining can effectively improve network system performance.