GPU parallel implementation of the approximate K-SVD algorithm using OpenCL. Irofti, P. & Dumitrescu, B. In 2014 22nd European Signal Processing Conference (EUSIPCO), pages 271-275, Sep., 2014.
![GPU parallel implementation of the approximate K-SVD algorithm using OpenCL [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper abstract bibtex
Paper abstract bibtex Training dictionaries for sparse representations is a time consuming task, due to the large size of the data involved and to the complexity of the training algorithms. We investigate a parallel version of the approximate K-SVD algorithm, where multiple atoms are updated simultaneously, and implement it using OpenCL, for execution on graphics processing units (GPU). This not only allows reducing the execution time with respect to the standard sequential version, but also gives dictionaries with which the training data are better approximated. We present numerical evidence supporting this somewhat surprising conclusion and discuss in detail several implementation choices and difficulties.
@InProceedings{6952053,
author = {P. Irofti and B. Dumitrescu},
booktitle = {2014 22nd European Signal Processing Conference (EUSIPCO)},
title = {GPU parallel implementation of the approximate K-SVD algorithm using OpenCL},
year = {2014},
pages = {271-275},
abstract = {Training dictionaries for sparse representations is a time consuming task, due to the large size of the data involved and to the complexity of the training algorithms. We investigate a parallel version of the approximate K-SVD algorithm, where multiple atoms are updated simultaneously, and implement it using OpenCL, for execution on graphics processing units (GPU). This not only allows reducing the execution time with respect to the standard sequential version, but also gives dictionaries with which the training data are better approximated. We present numerical evidence supporting this somewhat surprising conclusion and discuss in detail several implementation choices and difficulties.},
keywords = {approximation theory;graphics processing units;parallel programming;singular value decomposition;GPU parallel implementation;approximate K-SVD algorithm;OpenCL;training dictionaries;sparse representations;graphics processing units;Graphics processing units;Dictionaries;Matching pursuit algorithms;Sparse matrices;Kernel;Approximation algorithms;Parallel processing;sparse representation;dictionary design;parallel algorithm;GPU;OpenCL},
issn = {2076-1465},
month = {Sep.},
url = {https://www.eurasip.org/proceedings/eusipco/eusipco2014/html/papers/1569923277.pdf},
}Downloads: 0
{"_id":"iPHpXtuk8fuqdnthP","bibbaseid":"irofti-dumitrescu-gpuparallelimplementationoftheapproximateksvdalgorithmusingopencl-2014","authorIDs":[],"author_short":["Irofti, P.","Dumitrescu, B."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["P."],"propositions":[],"lastnames":["Irofti"],"suffixes":[]},{"firstnames":["B."],"propositions":[],"lastnames":["Dumitrescu"],"suffixes":[]}],"booktitle":"2014 22nd European Signal Processing Conference (EUSIPCO)","title":"GPU parallel implementation of the approximate K-SVD algorithm using OpenCL","year":"2014","pages":"271-275","abstract":"Training dictionaries for sparse representations is a time consuming task, due to the large size of the data involved and to the complexity of the training algorithms. We investigate a parallel version of the approximate K-SVD algorithm, where multiple atoms are updated simultaneously, and implement it using OpenCL, for execution on graphics processing units (GPU). This not only allows reducing the execution time with respect to the standard sequential version, but also gives dictionaries with which the training data are better approximated. We present numerical evidence supporting this somewhat surprising conclusion and discuss in detail several implementation choices and difficulties.","keywords":"approximation theory;graphics processing units;parallel programming;singular value decomposition;GPU parallel implementation;approximate K-SVD algorithm;OpenCL;training dictionaries;sparse representations;graphics processing units;Graphics processing units;Dictionaries;Matching pursuit algorithms;Sparse matrices;Kernel;Approximation algorithms;Parallel processing;sparse representation;dictionary design;parallel algorithm;GPU;OpenCL","issn":"2076-1465","month":"Sep.","url":"https://www.eurasip.org/proceedings/eusipco/eusipco2014/html/papers/1569923277.pdf","bibtex":"@InProceedings{6952053,\n author = {P. Irofti and B. Dumitrescu},\n booktitle = {2014 22nd European Signal Processing Conference (EUSIPCO)},\n title = {GPU parallel implementation of the approximate K-SVD algorithm using OpenCL},\n year = {2014},\n pages = {271-275},\n abstract = {Training dictionaries for sparse representations is a time consuming task, due to the large size of the data involved and to the complexity of the training algorithms. We investigate a parallel version of the approximate K-SVD algorithm, where multiple atoms are updated simultaneously, and implement it using OpenCL, for execution on graphics processing units (GPU). This not only allows reducing the execution time with respect to the standard sequential version, but also gives dictionaries with which the training data are better approximated. We present numerical evidence supporting this somewhat surprising conclusion and discuss in detail several implementation choices and difficulties.},\n keywords = {approximation theory;graphics processing units;parallel programming;singular value decomposition;GPU parallel implementation;approximate K-SVD algorithm;OpenCL;training dictionaries;sparse representations;graphics processing units;Graphics processing units;Dictionaries;Matching pursuit algorithms;Sparse matrices;Kernel;Approximation algorithms;Parallel processing;sparse representation;dictionary design;parallel algorithm;GPU;OpenCL},\n issn = {2076-1465},\n month = {Sep.},\n url = {https://www.eurasip.org/proceedings/eusipco/eusipco2014/html/papers/1569923277.pdf},\n}\n\n","author_short":["Irofti, P.","Dumitrescu, B."],"key":"6952053","id":"6952053","bibbaseid":"irofti-dumitrescu-gpuparallelimplementationoftheapproximateksvdalgorithmusingopencl-2014","role":"author","urls":{"Paper":"https://www.eurasip.org/proceedings/eusipco/eusipco2014/html/papers/1569923277.pdf"},"keyword":["approximation theory;graphics processing units;parallel programming;singular value decomposition;GPU parallel implementation;approximate K-SVD algorithm;OpenCL;training dictionaries;sparse representations;graphics processing units;Graphics processing units;Dictionaries;Matching pursuit algorithms;Sparse matrices;Kernel;Approximation algorithms;Parallel processing;sparse representation;dictionary design;parallel algorithm;GPU;OpenCL"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/Roznn/EUSIPCO/main/eusipco2014url.bib","creationDate":"2021-02-13T17:43:41.588Z","downloads":0,"keywords":["approximation theory;graphics processing units;parallel programming;singular value decomposition;gpu parallel implementation;approximate k-svd algorithm;opencl;training dictionaries;sparse representations;graphics processing units;graphics processing units;dictionaries;matching pursuit algorithms;sparse matrices;kernel;approximation algorithms;parallel processing;sparse representation;dictionary design;parallel algorithm;gpu;opencl"],"search_terms":["gpu","parallel","implementation","approximate","svd","algorithm","using","opencl","irofti","dumitrescu"],"title":"GPU parallel implementation of the approximate K-SVD algorithm using OpenCL","year":2014,"dataSources":["A2ezyFL6GG6na7bbs","oZFG3eQZPXnykPgnE"]}