First Experiences with Programming Persistent Data Structures

• Lucas Bastelli UNESP http://orcid.org/0000-0003-2201-0907
• Alexandro Baldassin UNESP https://orcid.org/0000-0001-8824-3055

DOI: https://doi.org/10.5753/eradsp.2022.222240

Abstract

Persistent memory is the latest storage technology. Unlike secundary storages devices (HD and SSD), its high access speed, lower latency and low granularity, allows direct connection with the processor's bus. But, to use this new technology, it is necessary new programming models to secure that data is always updated consistently. In this context, this paper shows the use of Intel's Persistent Memory Development Kit (PMDK) to build a singly linked list data structure. We discuss the challenges of programming for persistent rather than volatile memory and present the first results comparing the performance of commom DRAM and Intel Optane DC.

Keywords: Performance Measuremnt, Evaluation and Prediction, Programming Models, Memory Systems, Distributed Shared Memory (DSM)

References

Baldassin, A., Barreto, J., Castro, D., and Romano, P. (2021). Persistent memory: A survey of programming support and implementations. ACM Comput. Surv., 54(7).

Hennessy, J. L. and Patterson, D. A. (2017). Computer Architecture: A Quantitative Approach. Morgan Kaufmann, 6th edition.

Scargall, S. (2020). Programming Persistent Memory - A Comprehensive Guide for Developers. Apress, 1st edition.

Seltzer, M., Marathe, V., and Byan, S. (2018). An NVM Carol: Visions of NVM Past, Present, and Future. In Proceedings of the IEEE 34th International Conference on Data Engineering (ICDE), pages 15-23.