The Massive Memory Machine Project

海量记忆机器项目

• 批准号: 8420948
• 负责人: Richard Lipton
• 金额: $ 212.41万
• 依托单位: Princeton University
• 依托单位国家: 美国
• 项目类别: Cooperative Agreement
• 财政年份: 1985
• 资助国家: 美国
• 起止时间: 1985-07-15 至 1991-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8420948&HistoricalAwards=false
• 关键词: Massive; Memory; Machine; Project

This project is directed at the investigation of supercomputers with massive amounts of physical memory, in the tens of billions of bytes. The hypothesis of the research is that such machines may change, in fundamental ways, how certain classes of problems are solved and may lead to orders of magnitude performance improvements. Efforts are underway to validate experimentally the massive memory hypothesis, to explore ways to build such a machine in the future, and to investigate selected algorithmic processes for such designs. The validation portion of the project will be carried out on a Large Memory Computer (LMC), a conventional computer with an unconventional but economically feasible amount of memory (either 128 or 256 MB). The second part of the project will involve the study of a number of novel architectures for machines with billions of bytes of memory. A full-scale Massive Memory Machine prototype capable of supporting billions of bytes of memory will be constructed.

该项目旨在研究具有数百亿字节的海量物理内存的超级计算机。这项研究的假设是，这样的机器可能会从根本上改变某些问题的解决方式，并可能导致数量级的性能改进。人们正在努力通过实验验证大容量存储器假说，探索未来建造这样一台机器的方法，并研究为这种设计选择的算法过程。该项目的验证部分将在大型内存计算机（LMC）上进行，这是一种传统计算机，具有非常规但经济可行的内存量（128或256 MB）。该项目的第二部分将涉及研究具有数十亿字节内存的机器的许多新架构。一个能够支持数十亿字节内存的全尺寸大规模存储机原型将被构建。