Optimal compression of element pairs in fixed-width memories

Data compression is a well-studied (and well-solved) problem in the setup of long coding blocks. But important emerging applications need to compress data to memory words of small fixed widths. This new setup is the subject of this paper. In the problem we consider we have a source with a known disc...

Full description

Saved in:
Bibliographic Details
Published inITW : 2017 IEEE Information Theory Workshop : 6-10 November 2017 pp. 156 - 160
Main Authors Rottenstreich, Ori, Cassuto, Yuval
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.11.2017
Subjects
Conferences
Decoding
Dictionaries
Huffman coding
Memory management
Online AccessGet full text
DOI10.1109/ITW.2017.8278034

Cover

More Information
Summary:Data compression is a well-studied (and well-solved) problem in the setup of long coding blocks. But important emerging applications need to compress data to memory words of small fixed widths. This new setup is the subject of this paper. In the problem we consider we have a source with a known discrete distribution, and we wish to find a code that maximizes the success probability that two source instances can be represented together in L bits or less. A good practical use for this problem is a table with two-element entries that is stored in a memory of a fixed width L. Such tables of very large sizes are used in data-intensive computing applications. We solve the problem by efficiently finding an optimal code that uses a dictionary of linear size in the number of source elements.
DOI:10.1109/ITW.2017.8278034