CSR: Small: Collaborative Research: Instruction Address Translation Revisited

点击翻译按钮获取中文摘要

Applications continue to grow both in the size of their code and in the amountof data they access. Whether on mobile, desktop, or server platforms,applications are written in a modular fashion, relying heavily on commonfunctionality provided by a large number of libraries. Modern operatingsystems leverage this commonality by sharing libraries across applications toavoid duplicate copies of instructions in memory. However, operating systemsstill maintain independent memory management (address translation) datafor each process, consuming memory resources proportional to the number ofsimultaneously executing processes. The goal of this project is to enhancethe efficiency of memory system access and management for shared codelibraries. The success of these efforts will contribute to the long-termsustainability of the world's fast evolving digital economy and society,reducing computational waste and improving responsiveness. The educationcomponent will include improving classroom instruction by contributing newcourse modules as well as training graduate students in the importance ofefficient sharing.This project involves designing novel hardware and software to promoteefficient access to shared code in a multi-tasking environment, including: 1)new operating system memory management mechanisms and policies to eliminateduplication and minimize cache and translation lookaside buffer (TLB)footprint, examining the tradeoffs between the use of variable segments andhierarchical page tables; 2) reexamining the need for caching translationswhen utilizing emerging low-latency persistent memory; and 3) using availableprotection support for safe translation sharing in modern TLBs in addition todesigning new hardware support to improve efficiency.