A Scalable FPGA Engine for Parallel Acceleration of Singular Value Decomposition
Authors:
- 1 March 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 370-376
Abstract
Singular value decomposition (SVD) is a fundamental computational kernel and tool wildly used in data analytics such as least squares regression, principle components analysis (PCA), and pattern recognition. While a number of dedicated hardware processors have been proposed to accelerate the computationally intensive SVD computation, these designs suffer from poor flexibly and scalability, and/or lack full consideration of compute and data movement challenges associated with SVD. This paper presents a scalable parallel SVD FPGA engine based on the Hestenes-Jacobi method. We propose a so-called Maximum Data Sharing (MDS) ordering, which maximizes on-chip data reuse, and significantly reduces the expensive off-chip data movements and bandwidth requirement. Our SVD engine can flexibly decompose rectangular matrices with variable sizes and speed up SVD computation by $80\mathrm{X}$ to $300\mathrm{X}$ when compared with software SVD solvers such as the Eigen package running on high-performance CPUs. It can process much larger matrices than the previously reported FPGA designs.
Keywords
This publication has 11 references indexed in Scilit:
- Efficient one-sided Jacobi SVD computation on AMD GPU using OpenCL2016
- Local spectral variability features for speaker verificationDigital Signal Processing, 2016
- Real-Time Signal Processing of Massive Sensor Arrays via a Parallel Fast Converging SVD Algorithm: Latency, Throughput, and Resource AnalysisIEEE Sensors Journal, 2016
- An FPGA Implementation of the Hestenes-Jacobi Algorithm for Singular Value Decomposition2014
- Reconfigurable FPGA-Based Unit for Singular Value Decomposition of Large m x n Matrices2011
- Singular value decomposition utilizing parallel algorithms on graphical processors2011
- Improved SVD systolic array and implementation on FPGA2004
- Efficient implementation of Jacobi algorithms and Jacobi sets on distributed memory architecturesJournal of Parallel and Distributed Computing, 1990
- The Solution of Singular-Value and Symmetric Eigenvalue Problems on Multiprocessor ArraysSIAM Journal on Scientific and Statistical Computing, 1985
- Inversion of Matrices by Biorthogonalization and Related ResultsJournal of the Society for Industrial and Applied Mathematics, 1958