Performance Analysis of a Hybrid MPI/CUDA Implementation of the NAS-LU Benchmark

Pennycook, S.J., Hammond, S.D., Jarvis, S.A. and Mudalige, G.R. (2011) Performance Analysis of a Hybrid MPI/CUDA Implementation of the NAS-LU Benchmark. ACM SIGMETRICS Performance Evaluation Review, 38 (4). ISSN 0163-5999

[img]
Preview
 PDF - Submitted Version
Download (339Kb) | Preview

    Abstract

    We present the performance analysis of a port of the LU benchmark from the NAS Parallel Benchmark (NPB) suite to NVIDIA's Compute Unified Device Architecture (CUDA), and report on the optimisation efforts employed to take advantage of this platform. Execution times are reported for several different GPUs, ranging from low-end consumer-grade products to high-end HPC-grade devices, including the Tesla C2050 built on NVIDIA's Fermi processor.

    We also utilise recently developed performance models of LU to facilitate a comparison between future large-scale distributed clusters of GPU devices and existing clusters built on traditional CPU architectures, including a quad-socket, quad-core AMD Opteron cluster and an IBM BlueGene/P.

    Item Type: Article
    Uncontrolled Keywords: pcav hpsg performance cuda mpi lu hpc scientific solver bluegene performance modelling
    Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
    Divisions: Faculty of Science > Computer Science
    Depositing User: Simon Hammond
    Date Deposited: 01 Apr 2011 09:11
    Last Modified: 23 Feb 2012 09:07
    URI: http://eprints.dcs.warwick.ac.uk/id/eprint/634

    Actions (login required)

    		 View Item