LU Factorization with Partial Pivoting for a Multi-CPU, Multi-GPU Shared Memory System
Electrical Engineering and Computer Science, University of Tennessee
10th International Meeting on High-Performance Computing for Computational Science (VECPAR), 2012
@inproceedings{kurzak2012lu,
title={LU Factorization with Partial Pivoting for a Multi-CPU, Multi-GPU Shared Memory System},
author={Kurzak, Jakub and Luszczek, P and Faverge, Mathieu and Dongarra, Jack J and others},
booktitle={VECPAR 2012-10th International Meeting on High-Performance Computing for Computational Science},
year={2012}
}
LU factorization with partial pivoting is a canonical numerical procedure and the main component of the High Performance Linpack benchmark. This article presents an implementation of the algorithm for a hybrid, shared memory, system with standard CPU cores and GPU accelerators. The optimizations include lookahead, dynamic task scheduling, fine grain parallelism for memory-bound operations, autotuning, and data layout geared towards complex memory hierarchies. Performance in excess of one Tera flop/s is achieved using four AMD Magny Cours CPUs and four NVIDIA Fermi GPUs.
April 17, 2013 by hgpu
