SHF: Medium: Cross-Cutting Effort to Make Non-Volatile Memories Truly Usable

点击翻译按钮获取中文摘要

The emergence of non-volatile memory technologies offers great potential for improving the performance of modern software systems and applications, and is very likely to revolutionize multiple areas of computing. However, the use of non-volatile memory in computer systems faces substantial programming, performance, and usability challenges. These challenges hinder the widespread adoption of non-volatile memory in current and future computing systems, and make applications that use non-volatile memory difficult to write. To address these problems, this project uses a cross-cutting and multi-layer effort that makes advances in compilers, computer architecture, and distributed systems. With such an approach, the project significantly improves the software development productivity of programs using non-volatile memory technologies, as well as their performance. In addition, this project includes a robust education effort at the University of Illinois, where new interdisciplinary courses are developed and research opportunities are offered to undergraduate students and under-represented minorities.This project has three major thrusts. First, the project builds an easy-to-use and generic programming framework for non-volatile memories that requires minimal programmer involvement. Second, this project redesigns hardware primitives in the processor and memory hierarchy to minimize the overhead of both memory persistency operations and other operations needed to make the systems more usable. Third, this project develops distributed data persistency models and applies them to a distributed runtime environment that facilitates the use of non-volatile memories in distributed computing systems. The project further includes an effort to collaborate with Microsoft and Intel to perform technology transfer. Finally, this project maintains a public repository with all the data, code, results, emulators, and simulators used and developed throughout this project. The repository is accessible through http://iacoma.cs.uiuc.edu, and will be kept up to date for at least three years after the project ends.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.