The Design and Evaluation of a High-Performance Capability Mechanism for a VLSI Processor and its Impact on Software Reliability

VLSI处理器高性能能力机制的设计与评估及其对软件可靠性的影响

• 批准号: 8909512
• 负责人: Kanad Ghose
• 金额: $ 6万
• 依托单位: SUNY at Binghamton
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1989
• 资助国家: 美国
• 起止时间: 1989-08-01 至 1992-02-29
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8909512&HistoricalAwards=false
• 关键词: Design; Evaluation; Performance; Capability; Mechanism

A significant portion (70% and upwards, typically) of the cost of a medium to large-scale computing system is in its software. In turn, two-thirds or so of the software cost goes into maintenance--fixing the inevitable program bugs that are discovered later. Capability-based addressing has been suggested as a technique for detecting and avoiding many run-time errors because of its usefulness in enforcing small protection domains, data abstraction and secure module interfaces. Several machines proposed or implemented in the past to support capability-based addressing have been plagued by their abysmally slow speed. The purpose of the research project is to design a high-speed capability mechanism, examine and evaluate its implementation assuming realistic technological constraints. The project starts with the formulation of a simple, yet powerful set of primitives for the capability mechanism, followed by the design and simulation of an architecture using these mechanisms. High-level benchmarks would then be run on the simulator to evaluate the architecture.

一个中型到大型计算系统的很大一部分(通常是70%或以上)的成本都在其软件中。反过来，大约三分之二的软件成本用于维护--修复后来发现的不可避免的程序错误。基于能力的寻址被认为是一种检测和避免许多运行时错误的技术，因为它在执行小保护域、数据抽象和安全模块接口方面非常有用。过去提出或实现的支持基于能力的寻址的几台机器都受到速度极慢的困扰。研究项目的目的是设计一种高速能力机制，并在假设现实的技术约束下检查和评估其实施情况。该项目首先为能力机制制定一组简单但功能强大的原语，然后使用这些机制设计和模拟一个体系结构。然后，将在模拟器上运行高级别基准以评估体系结构。