Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A Primer on Compression in the Memory HierarchyMemory Compression

A Primer on Compression in the Memory Hierarchy: Memory Compression [While data compression can increase the utilization of caches and make better use of on-chip resources, applying it to main memory can potentially increase main memory capacity substantially at a low cost. This is important for many reasons. First of all, in server installations a substantial fraction of the infrastructure and ownership cost is due to main memory. Regarding cost of ownership, main memory consumes a lot of power, both dynamically as well as statically. Dynamically, a main memory access is several times more power consuming than an on-chip cache access. In addition, DRAM is volatile and needs to be refreshed. The refresh rate has steadily increased with the capacity of main memory, making it a significant component of the dynamic power consumption. As for static power consumption, it is proportional to the main memory capacity, which has also increased. Altogether, main memory accounts for a significant fraction of a system’s overall power consumption. Apart from component and cost of ownership of servers, other systems, and in particular mobile computers, are also sensitive to memory cost. In addition, they are also form-factor constrained; even if more memory were affordable, from a cost perspective, it may not be possible to fit more memory into a computerized device, e.g., a smart phone or a tablet.] http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png

A Primer on Compression in the Memory HierarchyMemory Compression

Download PDF
10 pages

Loading next page...
 
/lp/springer-journals/a-primer-on-compression-in-the-memory-hierarchy-memory-compression-LOXWGXYNWE
Publisher
Springer International Publishing
Copyright
© Springer Nature Switzerland AG 2016
ISBN
978-3-031-00623-4
Pages
33 –43
DOI
10.1007/978-3-031-01751-3_4
Publisher site
See Chapter on Publisher Site

Abstract

[While data compression can increase the utilization of caches and make better use of on-chip resources, applying it to main memory can potentially increase main memory capacity substantially at a low cost. This is important for many reasons. First of all, in server installations a substantial fraction of the infrastructure and ownership cost is due to main memory. Regarding cost of ownership, main memory consumes a lot of power, both dynamically as well as statically. Dynamically, a main memory access is several times more power consuming than an on-chip cache access. In addition, DRAM is volatile and needs to be refreshed. The refresh rate has steadily increased with the capacity of main memory, making it a significant component of the dynamic power consumption. As for static power consumption, it is proportional to the main memory capacity, which has also increased. Altogether, main memory accounts for a significant fraction of a system’s overall power consumption. Apart from component and cost of ownership of servers, other systems, and in particular mobile computers, are also sensitive to memory cost. In addition, they are also form-factor constrained; even if more memory were affordable, from a cost perspective, it may not be possible to fit more memory into a computerized device, e.g., a smart phone or a tablet.]

Published: Jan 1, 2016

There are no references for this article.