Solving Hyperbolic PDEs using Accelerator Architectures

Solving Hyperbolic PDEs using Accelerator Architectures

Show other versions (1)

Accelerator architectures are used to accelerate the simulation of nonlinear hyperbolic PDEs. Three different architectures, a multicore CPU using threading, IBM’s Cell Processor, and Nvidia’s Tesla GPUs are investigated. Speed-ups of between 40-75× relative to a single CPU core in single precisi...

Full description

Bibliographic Details
Main Author: Rostrup, Scott
Language:en
Published: 2009
Subjects:
GPU
Cell Processor
Hyperbolic PDEs
Hardware Optimization
Applied Mathematics
Online Access:http://hdl.handle.net/10012/4518
id ndltd-WATERLOO-oai-uwspace.uwaterloo.ca-10012-4518
record_format oai_dc
spelling ndltd-WATERLOO-oai-uwspace.uwaterloo.ca-10012-45182013-01-08T18:52:27ZRostrup, Scott2009-07-27T13:29:05Z2009-07-27T13:29:05Z2009-07-27T13:29:05Z2009-07-15http://hdl.handle.net/10012/4518Accelerator architectures are used to accelerate the simulation of nonlinear hyperbolic PDEs. Three different architectures, a multicore CPU using threading, IBM’s Cell Processor, and Nvidia’s Tesla GPUs are investigated. Speed-ups of between 40-75× relative to a single CPU core in single precision are obtained using the Cell processor and the GPU. The three implementations are extended to parallel computing clusters by making use of the Message Passing Interface (MPI). The resulting hybrid-parallel code is investigated for performance and scalability on both a GPU and Cell computing cluster.enGPUCell ProcessorHyperbolic PDEsHardware OptimizationSolving Hyperbolic PDEs using Accelerator ArchitecturesThesis or DissertationApplied MathematicsMaster of MathematicsApplied Mathematics
collection NDLTD
language en
sources NDLTD
topic GPU
Cell Processor
Hyperbolic PDEs
Hardware Optimization
Applied Mathematics
spellingShingle GPU
Cell Processor
Hyperbolic PDEs
Hardware Optimization
Applied Mathematics
Rostrup, Scott
Solving Hyperbolic PDEs using Accelerator Architectures
description Accelerator architectures are used to accelerate the simulation of nonlinear hyperbolic PDEs. Three different architectures, a multicore CPU using threading, IBM’s Cell Processor, and Nvidia’s Tesla GPUs are investigated. Speed-ups of between 40-75× relative to a single CPU core in single precision are obtained using the Cell processor and the GPU. The three implementations are extended to parallel computing clusters by making use of the Message Passing Interface (MPI). The resulting hybrid-parallel code is investigated for performance and scalability on both a GPU and Cell computing cluster.
author Rostrup, Scott
author_facet Rostrup, Scott
author_sort Rostrup, Scott
title Solving Hyperbolic PDEs using Accelerator Architectures
title_short Solving Hyperbolic PDEs using Accelerator Architectures
title_full Solving Hyperbolic PDEs using Accelerator Architectures
title_fullStr Solving Hyperbolic PDEs using Accelerator Architectures
title_full_unstemmed Solving Hyperbolic PDEs using Accelerator Architectures
title_sort solving hyperbolic pdes using accelerator architectures
publishDate 2009
url http://hdl.handle.net/10012/4518
work_keys_str_mv AT rostrupscott solvinghyperbolicpdesusingacceleratorarchitectures
_version_ 1716573354946723840

Similar Items