CUDAICA: GPU optimization of infomax-ICA EEG analysis

In recent years, Independent Component Analysis (ICA) has become a standard to identify relevant dimensions of the data in neuroscience. ICA is a very reliable method to analyze data but it is, computationally, very costly. The use of ICA for online analysis of the data, used in brain computing inte...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Kamienkowski, Juan Esteban, Sigman, Mariano, Fernández Slezak, Diego
Publicado: 2012
Materias:
Brain computing
Computing time
EEG analysis
Function calls
Independent signals
Multiple channels
On-line analysis
Vector processors
Video cards
Brain computer interface
Matrix algebra
Independent component analysis
algorithm
article
computer graphics
computer program
electroencephalography
human
Internet
signal processing
Algorithms
Computer Graphics
Electroencephalography
Humans
Signal Processing, Computer-Assisted
Software
Acceso en línea:https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_16875265_v2012_n_p_Raimondo
http://hdl.handle.net/20.500.12110/paper_16875265_v2012_n_p_Raimondo
Aporte de:
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) de Universidad de Buenos Aires
id paper:paper_16875265_v2012_n_p_Raimondo
record_format dspace
spelling paper:paper_16875265_v2012_n_p_Raimondo2023-06-08T16:26:40Z CUDAICA: GPU optimization of infomax-ICA EEG analysis Kamienkowski, Juan Esteban Sigman, Mariano Fernández Slezak, Diego Brain computing Computing time EEG analysis Function calls Independent signals Multiple channels On-line analysis Vector processors Video cards Brain computer interface Matrix algebra Independent component analysis algorithm article computer graphics computer program electroencephalography human Internet signal processing Algorithms Computer Graphics Electroencephalography Humans Internet Signal Processing, Computer-Assisted Software In recent years, Independent Component Analysis (ICA) has become a standard to identify relevant dimensions of the data in neuroscience. ICA is a very reliable method to analyze data but it is, computationally, very costly. The use of ICA for online analysis of the data, used in brain computing interfaces, results are almost completely prohibitive. We show an increase with almost no cost (a rapid video card) of speed of ICA by about 25 fold. The EEG data, which is a repetition of many independent signals in multiple channels, is very suitable for processing using the vector processors included in the graphical units. We profiled the implementation of this algorithm and detected two main types of operations responsible of the processing bottleneck and taking almost 80 of computing time: vector-matrix and matrix-matrix multiplications. By replacing function calls to basic linear algebra functions to the standard CUBLAS routines provided by GPU manufacturers, it does not increase performance due to CUDA kernel launch overhead. Instead, we developed a GPU-based solution that, comparing with the original BLAS and CUBLAS versions, obtains a 25x increase of performance for the ICA calculation. © Copyright 2012 Federico Raimondo et al. Fil:Kamienkowski, J.E. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales; Argentina. Fil:Sigman, M. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales; Argentina. Fil:Fernandez Slezak, D. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales; Argentina. 2012 https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_16875265_v2012_n_p_Raimondo http://hdl.handle.net/20.500.12110/paper_16875265_v2012_n_p_Raimondo
institution Universidad de Buenos Aires
institution_str I-28
repository_str R-134
collection Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA)
topic Brain computing
Computing time
EEG analysis
Function calls
Independent signals
Multiple channels
On-line analysis
Vector processors
Video cards
Brain computer interface
Matrix algebra
Independent component analysis
algorithm
article
computer graphics
computer program
electroencephalography
human
Internet
signal processing
Algorithms
Computer Graphics
Electroencephalography
Humans
Internet
Signal Processing, Computer-Assisted
Software
spellingShingle Brain computing
Computing time
EEG analysis
Function calls
Independent signals
Multiple channels
On-line analysis
Vector processors
Video cards
Brain computer interface
Matrix algebra
Independent component analysis
algorithm
article
computer graphics
computer program
electroencephalography
human
Internet
signal processing
Algorithms
Computer Graphics
Electroencephalography
Humans
Internet
Signal Processing, Computer-Assisted
Software
Kamienkowski, Juan Esteban
Sigman, Mariano
Fernández Slezak, Diego
CUDAICA: GPU optimization of infomax-ICA EEG analysis
topic_facet Brain computing
Computing time
EEG analysis
Function calls
Independent signals
Multiple channels
On-line analysis
Vector processors
Video cards
Brain computer interface
Matrix algebra
Independent component analysis
algorithm
article
computer graphics
computer program
electroencephalography
human
Internet
signal processing
Algorithms
Computer Graphics
Electroencephalography
Humans
Internet
Signal Processing, Computer-Assisted
Software
description In recent years, Independent Component Analysis (ICA) has become a standard to identify relevant dimensions of the data in neuroscience. ICA is a very reliable method to analyze data but it is, computationally, very costly. The use of ICA for online analysis of the data, used in brain computing interfaces, results are almost completely prohibitive. We show an increase with almost no cost (a rapid video card) of speed of ICA by about 25 fold. The EEG data, which is a repetition of many independent signals in multiple channels, is very suitable for processing using the vector processors included in the graphical units. We profiled the implementation of this algorithm and detected two main types of operations responsible of the processing bottleneck and taking almost 80 of computing time: vector-matrix and matrix-matrix multiplications. By replacing function calls to basic linear algebra functions to the standard CUBLAS routines provided by GPU manufacturers, it does not increase performance due to CUDA kernel launch overhead. Instead, we developed a GPU-based solution that, comparing with the original BLAS and CUBLAS versions, obtains a 25x increase of performance for the ICA calculation. © Copyright 2012 Federico Raimondo et al.
author Kamienkowski, Juan Esteban
Sigman, Mariano
Fernández Slezak, Diego
author_facet Kamienkowski, Juan Esteban
Sigman, Mariano
Fernández Slezak, Diego
author_sort Kamienkowski, Juan Esteban
title CUDAICA: GPU optimization of infomax-ICA EEG analysis
title_short CUDAICA: GPU optimization of infomax-ICA EEG analysis
title_full CUDAICA: GPU optimization of infomax-ICA EEG analysis
title_fullStr CUDAICA: GPU optimization of infomax-ICA EEG analysis
title_full_unstemmed CUDAICA: GPU optimization of infomax-ICA EEG analysis
title_sort cudaica: gpu optimization of infomax-ica eeg analysis
publishDate 2012
url https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_16875265_v2012_n_p_Raimondo
http://hdl.handle.net/20.500.12110/paper_16875265_v2012_n_p_Raimondo
work_keys_str_mv AT kamienkowskijuanesteban cudaicagpuoptimizationofinfomaxicaeeganalysis
AT sigmanmariano cudaicagpuoptimizationofinfomaxicaeeganalysis
AT fernandezslezakdiego cudaicagpuoptimizationofinfomaxicaeeganalysis
_version_ 1768541864996110336

Ejemplares similares