On the performance of BWT sorting algorithms

Seward, John P.

doi:10.1109/dcc.2000.838157

Cited by 53 publications

(43 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Seward [18] shows the complexity of this algorithm to be O(A · n log n), where A is the average number of symbols per rotation that must be compared to verify that the rotations are in order.…”

Section: Burrows-wheeler Transformmentioning

confidence: 99%

Parallel lossless data compression on the GPU

Patel

Zhang

Mak

et al. 2012

2012 Innovative Parallel Computing (InPar)

View full text Add to dashboard Cite

We present parallel algorithms and implementations of a bzip2-like lossless data compression scheme for GPU architectures. Our approach parallelizes three main stages in the bzip2 compression pipeline: Burrows-Wheeler transform (BWT), move-to-front transform (MTF), and Huffman coding. In particular, we utilize a two-level hierarchical sort for BWT, design a novel scan-based parallel MTF algorithm, and implement a parallel reduction scheme to build the Huffman tree. For each algorithm, we perform detailed performance analysis, discuss its strengths and weaknesses, and suggest future directions for improvements. Overall, our GPU implementation is dominated by BWT performance and is 2.78× slower than bzip2, with BWT and MTFHuffman respectively 2.89× and 1.34× slower on average.

show abstract

Section: Burrows-wheeler Transformmentioning

confidence: 99%

Parallel lossless data compression on the GPU

Patel

Zhang

Mak

et al. 2012

2012 Innovative Parallel Computing (InPar)

View full text Add to dashboard Cite

show abstract

“…Star transform [14], length index preserving transformation (LIPT) [15], star new transform (StarNT) [16], and word replacement transformation (WRT) [17] are preprocessing methods that use a static dictionary. Other algorithms that cannot be classified in either of the two groups are run-length encoders (RLE) [18], Burrows-Wheeler transform (BWT) ( [19,20]), and BZIP2 [21], which uses block sorting BWT together with Huffman coding.…”

Section: Introductionmentioning

confidence: 99%

A new dictionary-based preprocessor that uses radix-190 numbering

Şenergin¹,

Ince

2016

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

Abstract:Various scholarly works in the literature have pointed out that placing a preprocessor in front of a standard postcompressor would help achieve higher gains while compressing natural-language text files. Ever since, there has been much research on preprocessors to improve the gain attained by concatenated systems. With the same goal in mind our paper proposes a new word-based preprocessor named METEHAN190 (M190) and contrasts its performance with four other state-of-the-art preprocessors. Throughout the experiments source files from the Wall Street Journal (WSJ) archive, and the Calgary, Canterbury, Gutenberg, and Pizza and Chili corpora were used. Postcompressors adapted were Prediction by Partial Matching compressor using method-D (PPMD) and Monstrous PPM II compressor (PPMonstr).It was observed that in all three experiments WRT and M190 would achieve the two highest compression gains. For small text and transcription files from the Calgary corpus, M190 would outperform all preprocessors including WRT.On the other hand, a look at average encoding and decoding times shows that the semistatic byte-oriented methods are much faster in comparison to the static dictionary-based methods that encode words with characters.

show abstract

“…The compression results also show an improvement in the range of around 2% to 7% for compression methods with LIPT over the best results obtained by various modifications in BWT [1,2,5,18] and PPM [6,7,17,19]. This comes at the expense of some storage overhead whose amortized cost is shown to be negligible.…”

Section: Figure 1: Frequency Of Words Versus Length Of Words In Our Tmentioning

confidence: 55%

“…PPMD is a variant of PPM which computes escape probability as (u/2)/n, where u is the number of unique token seen so far, and n is the number of token seen so far [11]. BWT has proved to be the most efficient and a number of efforts have been made to improve its efficiency [1,2,5,18]. PPM gives better compression ratio than that of BWT but is very slow.…”

Section: Introductionmentioning

confidence: 99%