CISE-MSI : RCBP-ED: CCF-FET : Improving Reliability and Durability in Phase Change Main Memory (PCM)

Cal Poly Humboldt Sponsored Programs Foundation, Arcata CA

Investigators

Abstract

Phase Change Memories (PCMs) are a cutting-edge type of computer memory technology that can revolutionize data storage and retrieval. PCMs use special materials that can switch between two states, similar to how ice can change from solid to liquid with temperature. This unique property allows PCMs to store and retrieve data quickly, even when the power is turned off. However, PCMs face challenges related to their endurance and reliability. Researchers are working to improve the durability and reliability of PCMs to make them more robust and suitable for various applications. The significance of PCM research extends and has the potential to bridge the performance gap between storage and main memory, leading to faster and more efficient computing systems. This can have a profound impact on fields such as data analysis, artificial intelligence, and scientific research by enabling faster and more reliable data access. Investing in PCM research through organizations like NSF is vital because this research has the potential to shape the future of computing, leading to faster and more energy-efficient systems with transformative applications that benefit society. This project aims to explore efficient computer algorithms and architecture for the design and implementation of PCM to fulfill the storage requirements of consumers. The main focus is on enhancing the reliability and durability of PCM technology, which serves as a crucial technology for non-volatile electrical data storage at the nanometer scale. This project tackles problems from both hardware and software perspectives. On the hardware side, all proposed solutions strive to improve the performance of a PCM main memory system while addressing concerns related to limited endurance, hard errors, and drift-induced soft errors. From a software standpoint, it involves operating system support for address translation from virtual memory addresses to physical addresses and from physical addresses to actual PCM addresses. 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.