Year: 2020
Type: book-chapter
Abstract: Taking advantage of the unique characteristics of persistent memory, such as byte addressability, persistence, and update in place, allows us to build data structures that are much faster than any data structure requiring serialization or flushing to a disk. However, this comes at a cost. Algorithms must be carefully designed to properly persist data by flushing CPU caches or using non-temporal stores and memory barriers to maintain data consistency. This chapter describes how to design such data structures and algorithms and shows what properties they should have.
Source: Apress eBooks
Author: Steve Scargall
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Funder: Intel Corporation
OpenAlex ID: W2998750529