一个基于IA-64体系的内存管理大页面的实现模型

A Model of Realization of Large Page of Memory Management Based on IA-64

本文提出了一种基于IA-64体系结构的内存页面大页面化的模型,可执行文件ELF的Data Segment使用大页面。由于转换解析缓冲区(TLB)能映射更大的虚拟内存范围,从而可减小未命中率,因此可以提高使用大页面的高性能计算(HPC)应用程序或使用大量虚拟内存的任何内存访问密集型应用程序系统性能。

A model of large page of memory management is put forward in this article, and the Data Segment of the ELF file uses large page of memory. Because when large page is used, TLB（translation lookaside buffer） can map much larger virtual memory area, which leads to the reduction of miss rate. Therefore, computing performance of HPC applications or any other applications which use a large amount of virtual memory and access the memory extremely frequently will be improved.