Title :
A work-efficient parallel algorithm for constructing Huffman codes
Author :
Milidiú, Ruy Luiz ; Laber, Eduardo Sany ; Pessoa, Artur Rives
Author_Institution :
Dept. de Inf., PUC-Rio, Rio de Janeiro, Brazil
Abstract :
Given an alphabet Σ={a1,...,an) and a corresponding list of weights [w1,...,wn], a Huffman code for this alphabet is a prefix code that minimizes the weighted length of a code string, defined to be Σi=1 nwili, where li is the length of the code assigned to ai. We present ES-ParHuff, a work-efficient PRAM CREW algorithm for constructing Huffman codes. An important feature of the algorithm is its simplicity. This algorithm is a direct parallelization of Huffman´s algorithm. ES-ParHuff runs in O(Hloglog(n/H)) time with O(n) work, where H is the length of the longest generated code
Keywords :
Huffman codes; computational complexity; minimisation; parallel algorithms; ES-ParHuff; Huffman codes; PRAM CREW algorithm; code string; prefix code; running time; weighted length minimization; work-efficient parallel algorithm; Binary trees; Chromium; Huffman coding; Merging; Parallel algorithms; Phase change random access memory;
Conference_Titel :
Data Compression Conference, 1999. Proceedings. DCC '99
Conference_Location :
Snowbird, UT
Print_ISBN :
0-7695-0096-X
DOI :
10.1109/DCC.1999.755677