scispace - formally typeset
Search or ask a question
Proceedings Article•DOI•

Design of a language-independent parallel string matching unit for NLP

V.S. Murty1, P. C. Reghu Raj1, S. Raman1•
Indian Institute of Technology Madras1
12 May 2003-pp 149-155
TL;DR: The FPGA design of a dedicated hardware for string matching that uses memory interleaving and parallel processing techniques can relieve the host CPU from this burden, thereby making the system suitable for real-time applications.
Abstract: In natural language processing applications, string matching is the main time-consuming operation due to the large size of lexicon. Data dependence is minimal in string matching operations, and hence it is ideal for parallelization. A dedicated hardware for string matching that uses memory interleaving and parallel processing techniques can relieve the host CPU from this burden, thereby making the system suitable for real-time applications. This paper reports the FPGA design of such a system with m parallel matching units. The time complexity of the proposed algorithm is O (log2 n), where n is the total number of lexical entries. This has been achieved by a proper selection of the value of m. A special memory organization technique, which reduces the storage space by nearly 70%, has been adopted for storing lexical entries. The techniques used for matching and storage of lexical entries make the system language independent
Citations
More filters
Journal Article•DOI•
Performance Analysis of GPU V/S CPU for Image Processing Applications

[...]

B. N. Manjunatha Reddy
28 Feb 2017-International Journal for Research in Applied Science and Engineering Technology
TL;DR: The research results shows that the GPU implementation can achieve a speed up of more than 60% of time in comparison with CPU implementation of image processing, indicating GPU has great potential as high-performance co-processor.
Abstract: Image processing grosses much more time to perform the convolution in image filtering on CPU, since the computation demand of image filtering is enormous. Contrast to CPU, Graphics Processing Unit (GPU)is a good way to accelerate the image processing. By comparison and analysis,it has reached a conclusion that GPU is appropriate for processinglarge-scale data-parallel load of high-density computing.CUDA(Compute Unified Device Architecture) is a parallel computing architecture established by NVIDIA. CUDA is highly suitable for general purpose programming on GPU which is a programming interface to use the parallel architecture for general purpose computing. This paper stressesthe possible gain in time which can be attained on comparison and analysis of GPU over CPU implementation and the research results shows that the GPU implementation can achieve a speed up of more than 60%of time in comparison with CPU implementation of image processing. GPU has great potential as high-performance co-processor.

7 citations

References
More filters
Book•
Natural Language Understanding 2nd Ed

[...]

James F. Allen1•
University of Rochester1
01 Jan 2002

388 citations

Book•
Efficient parsing for natural language

[...]

Masaru Tomita
01 Jan 1985
TL;DR: One of recommendation of the book that you need to read is shown, which is a kind of precious book written by an experienced author and it will show the reasonable reasons why you should read this book.
Abstract: Any books that you read, no matter how you got the sentences that have been read from the books, surely they will give you goodness. But, we will show you one of recommendation of the book that you need to read. This efficient parsing for natural language is what we surely mean. We will show you the reasonable reasons why you need to read this book. This book is a kind of precious book written by an experienced author.

387 citations

"Design of a language-independent pa..." refers background in this paper

  • ...Method 1: Storing all inflections in the lexicon [8]....

    [...]

Journal Article•DOI•
Data structures and algorithms for approximate string matching

[...]

Zvi Galil1, Raffaele Giancarlo1•
Columbia University1
01 Mar 1988-Journal of Complexity
TL;DR: This paper surveys techniques for designing efficient sequential and parallel approximate string matching algorithms and special attention is given to the methods for the construction of data structures that efficiently support primitive operations needed in approximatestring matching.

153 citations

Journal Article•DOI•
ASIC design of a matching unit for NLP

[...]

S. Raman1, E.R. Shaji1•
Indian Institute of Technology Madras1
01 Aug 1995-Microprocessors and Microsystems
TL;DR: The logic verification and timing analysis of the matching unit were carried out using the Actel Viewlogic System, after which the FPGA devices were fused.

5 citations

"Design of a language-independent pa..." refers background or methods in this paper

  • .../0 signals of the functional unit 0-7803-7970-5/03/$20.00 ©2003 IEEE Functional Unit Table 3 Comparative space saving with different lexicon organization methods Method2) 1 0.25 0.25 0.25 0.25 53.1 64.2 27.5 2 0.6 0.2 0.15 0.05 53.1 70.6 32.96 3 0.25 0.5 0.125 0.125 53.1 66.9 29.3 4 0.125 0.125 0.5 0.25 53.1 61.5 17.8 5 0.125 0.125 0.25 0.5 53.1 59.6 13.8 Average space saving there will not be any match between input word and the lexical entry....

    [...]

  • ...Depending on the lengths of the two words under comparison, and their characters in the respective positions, three parameters, namely, VALUE, LEVEL and LCI (Last Character Information) are computed for determining the type of match [7]....

    [...]

  • ...Method 1 [61 1 ((n*k)+ 1) /I (n*k) Method2 [7] 1 (n+1)/2 N Method3 [ours] I (n/(j *s)+ 1)/2 (n/(j*s))...

    [...]

  • ...Shaji and Raman [7] describe a linear algorithm that stores extra information on the inflections of a word in the lexicon to aid the process of matching....

    [...]

  • ...There are two kinds of matches for an input word with the lexical entries: the Perfect Match (PM) and the Approximate Match (AM) [7]....

    [...]

Proceedings Article•DOI•
Design of a high speed string matching co-processor for NLP

[...]

V.S. Murty1, P. C. Reghu Raj1, S. Raman1•
Indian Institute of Technology Madras1
04 Jan 2003
TL;DR: The FPGA design of a system with m parallel matching units is reported, shown to improve the performance by a factor of nearly m, without increasing the chip area by more than 45%
Abstract: In Natural Language Processing applications, string matching is the main time-consuming operation A dedicated co-processor for string matching that uses memory interleaving and parallel processing techniques can relieve the host CPU from this burden This paper reports the FPGA design of such a system with m parallel matching units It has been shown to improve the performance by a factor of nearly m, without increasing the chip area by more than 45% The time complexity of the proposed algorithm is O(log/sub 2/ n), where n is the number of lexical entries The memory used by the lexicon has been efficiently organized and the space saving achieved is about 67%

5 citations

"Design of a language-independent pa..." refers background in this paper

  • ...Based on the details of the calculation of memory space saving discussed elsewhere [9], Table 3 entries were arrived at....

    [...]

Related Papers (5)
Families of FPGA-based algorithms for approximate string matching

[...]

27 Sep 2004
T. Van Court, Martin C. Herbordt
Bulk Processing in Distributed Logic Memory

[...]

01 Apr 1965-IEEE Transactions on Electronic Computers
B. A. Crane, J. A. Githens
A Parallel “String Matching Engine” for use in High Speed Network Intrusion Detection Systems

[...]

01 Jun 2006-Journal in Computer Virology
Gerald Tripp
Concurrent Processing Memory

[...]

05 Jun 2004-arXiv: Distributed, Parallel, and Cluster Computing
Chengpu Wang
A high performance architecture for computing burrows-wheeler transform on FPGAs

[...]

01 Jan 2013
Umer I. Cheema, Ashfaq Khokhar