AI Chat Paper
Note: Please note that the following content is generated by AMiner AI. SciOpen does not take any responsibility related to this content.
{{lang === 'zh_CN' ? '文章概述' : 'Summary'}}
{{lang === 'en_US' ? '中' : 'Eng'}}
Chat more with AI
Powered by AMiner. Clicking this button will take you away from SciOpen.
Home Journal of Computer Science and Technology Article
Article Link
Cite
EndNote(RIS) BibTeX
Collect
Submit Manuscript
Show Outline
Outline
Show full outline
Hide outline
Outline
Show full outline
Hide outline
Regular Paper

Usage Scenarios for Byte-Addressable Persistent Memory in High-Performance and Data Intensive Computing

EPCC, The University of Edinburgh, Edinburgh EH8 9BT, U.K.
System Vertrieb Alexander GmbH, Paderborn 33106, Germany
Show Author Information

Abstract

Byte-addressable persistent memory (B-APM) presents a new opportunity to bridge the performance gap between main memory and storage. In this paper, we present the usage scenarios for this new technology, based on the capabilities of Intel’s DCPMM. We outline some of the basic performance characteristics of DCPMM, and explain how it can be configured and used to address the needs of memory and I/O intensive applications in the HPC (high-performance computing) and data intensive domains. Two decision trees are presented to advise on the configuration options for B-APM; their use is illustrated with two examples. We show that the flexibility of the technology has the potential to be truly disruptive, not only because of the performance improvements it can deliver, but also because it allows systems to cater for wider range of applications on homogeneous hardware.

Keywords

byte-addressabledata intensivememory intensivepersistent

Electronic Supplementary Material

Download File(s)
jcst-36-1-110-Highlights.pdf (233.1 KB)

References

[1]
Jackson A, Weiland M, Parsons M, Homölle B. An architecture for high performance computing and data systems using byte-addressable persistent memory. In Proc. the 2019 ISC High Performance International Workshops, June 2019, pp.258-274. DOI: 10.1007/978-3-030-34356-9_21.
Crossref
[2]
Izraelevitz J, Yang J, Zhang L et al. Basic performance measurements of the Intel Optane DC persistent memory module. arXiv: 1903.05714, 2019. http://arxiv.org/abs/1903.05714, March 2020.
[3]
Patil O, Ionkov L, Lee J et al. Performance characterization of a DRAM-NVM hybrid memory architecture for HPC applications using Intel Optane DC persistent memory modules. In Proc. the International Symposium on Memory Systems, September 2019, pp.288-303. DOI: 10.1145/3357526.3357541.
Crossref
[4]

Mason T, Doudali T D, Seltzer M, Gavrilovska A. Unexpected performance of Intel® OptaneTM DC persistent memory. IEEE Computer Architecture Letters, 2020, 19(1): 55-58. DOI: 10.1109/LCA.2020.2987303.
Crossref Google Scholar
[5]

Clark S J, Segall M D, Pickard C J et al. First principles methods using CASTEP. Zeitschrift für Kristallographie, 2005, 220: 567-570. DOI: 10.1524/zkri.220.5.567.65075.
Crossref Google Scholar
[6]
Weiland M, Brunst H, Quintino T et al. An early evaluation of Intel’s Optane DC persistent memory module and its impact on high-performance scientific applications. In Proc. the International Conference for High Performance Computing, Networking, Storage and Analysis, Nov. 2019, Article No. 76. DOI: 10.1145/3295500.3356159.
Crossref
[7]

Vef M A, Moti N, Süß T et al. GekkoFS—A temporary burst buffer file system for HPC applications. J. Comput. Sci. Technol., 2020, 35(1): 72-91. DOI: 10.1007/s11390-020-9797-6.
Crossref Google Scholar
[8]

Brinkmann A, Mohror K, Yu W et al. Ad hoc file systems for high-performance computing. J. Comput. Sci. Technol., 2020, 35(1): 4-26. DOI: 10.1007/s11390-020-9801-1.
Crossref Google Scholar
[9]
Smart S, Quintino T, Raoult B. A scalable object store for meteorological and climate data. In Proc. the Platform for Advanced Scientific Computing Conference, June 2017, Article No. 13. DOI: 10.1145/3093172.3093238.
Crossref
[10]
Smart S, Quintino T, Raoult B. A high-performance distributed object-store for exascale numerical weather prediction and climate. In Proc. the Platform for Advanced Scientific Computing Conference, June 2019, Article No. 16. DOI: 10.1145/3324989.3325726.
Crossref
[11]

Weiland M, Jackson A, Johnson N, Parsons M. Exploiting the performance benefits of storage class memory for HPC and HPDA workows. Journal of Supercomputing Frontiers and Innovations, 2018, 5(1): 79-94. DOI: 10.14529/jsfi180105.
Crossref Google Scholar
[12]
Miranda A, Jackson A, Tocci T, Panourgias I, Nou R. NORNS: Extending slurm to support data-driven workflows through asynchronous data staging. In Proc. the 2019 IEEE International Conference on Cluster Computing, Sept. 2019. DOI: 10.1109/CLUSTER.2019.8891014.
Crossref
[13]
Brown N, Weiland M, Hill A et al. A highly scalable Met Office NERC Cloud model. In Proc. the 3rd International Conference on Exascale Applications and Software, April 2015, pp.132-137. DOI: 10.5555/2820083.2820108.
Journal of Computer Science and Technology
Volume 36 Issue 1,
January 2021
Pages 110-122
DOI: 10.1007/s11390-020-0776-8
Cite this article:
Weiland M, Homölle B. Usage Scenarios for Byte-Addressable Persistent Memory in High-Performance and Data Intensive Computing. Journal of Computer Science and Technology, 2021, 36(1): 110-122. https://doi.org/10.1007/s11390-020-0776-8

359

Views

2

Crossref

1

Web of Science

2

Scopus

0

CSCD

Altmetrics
Received: 03 July 2020
Accepted: 09 December 2020
Published: 05 January 2021
© Institute of Computing Technology, Chinese Academy of Sciences 2021
Return