A Practical GPU-Accelerated Implementation of Orthogonal Matching Pursuit

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lubonja, Ariel, Praesius, Sebastian Kazmarek, Tran, Trac Duy
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912989432512512
author Lubonja, Ariel
Praesius, Sebastian Kazmarek
Tran, Trac Duy
author_facet Lubonja, Ariel
Praesius, Sebastian Kazmarek
Tran, Trac Duy
contents Finding the sparsest solution to the underdetermined system $\mathbf{y}=\mathbf{Ax}$, given a tolerance, is known to be NP-hard. Many approximate solutions to this problem exist, and Orthogonal Matching Pursuit (OMP) is one of the most widely used. However, existing OMP implementations don't take full advantage of matrix properties or modern CPU and GPU-based Linear Algebra kernels. For this paper, we implemented an efficient implementation of OMP that leverages Cholesky inverse properties as well as the power of GPUs to deliver up to \textbf{310x speedup over Scikit-Learn} and \textbf{26x over SPAMS}. The package is published on PyPI (\texttt{pip install batched-omp}) and is fully scikit-learn compatible.
format Preprint
id arxiv_https___arxiv_org_abs_2407_06434
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Practical GPU-Accelerated Implementation of Orthogonal Matching Pursuit
Lubonja, Ariel
Praesius, Sebastian Kazmarek
Tran, Trac Duy
Distributed, Parallel, and Cluster Computing
Finding the sparsest solution to the underdetermined system $\mathbf{y}=\mathbf{Ax}$, given a tolerance, is known to be NP-hard. Many approximate solutions to this problem exist, and Orthogonal Matching Pursuit (OMP) is one of the most widely used. However, existing OMP implementations don't take full advantage of matrix properties or modern CPU and GPU-based Linear Algebra kernels. For this paper, we implemented an efficient implementation of OMP that leverages Cholesky inverse properties as well as the power of GPUs to deliver up to \textbf{310x speedup over Scikit-Learn} and \textbf{26x over SPAMS}. The package is published on PyPI (\texttt{pip install batched-omp}) and is fully scikit-learn compatible.
title A Practical GPU-Accelerated Implementation of Orthogonal Matching Pursuit
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2407.06434