Finite element integration on GPUs

Matthew G. Knepley; Andy R. Terrel

doi:10.1145/2427023.2427027

Finite element integration on GPUs

Matthew G. Knepley
, Andy R. Terrel

Department of Computer Science and Engineering

University of Texas at Austin

Research output: Contribution to journal › Article › peer-review

30 Scopus citations

Abstract

We present a novel finite element integration method for low-order elements on GPUs. We achieve more than 100GF for element integration on first order discretizations of both the Laplacian and Elasticity operators on an NVIDIA GTX285, which has a nominal single precision peak flop rate of 1 TF/s and bandwidth of 159 GB/s, corresponding to a bandwidth limited peak of 40 GF/s.

Original language	English
Article number	10
Journal	ACM Transactions on Mathematical Software
Volume	39
Issue number	2
DOIs	https://doi.org/10.1145/2427023.2427027
State	Published - Feb 2013

Keywords

Finite element integration
GPGPUs

Access to Document

10.1145/2427023.2427027

Cite this

@article{bdb40806cd7a4b1bbcc982d0c3967517,

title = "Finite element integration on GPUs",

abstract = "We present a novel finite element integration method for low-order elements on GPUs. We achieve more than 100GF for element integration on first order discretizations of both the Laplacian and Elasticity operators on an NVIDIA GTX285, which has a nominal single precision peak flop rate of 1 TF/s and bandwidth of 159 GB/s, corresponding to a bandwidth limited peak of 40 GF/s.",

keywords = "Finite element integration, GPGPUs",

author = "Knepley, \{Matthew G.\} and Terrel, \{Andy R.\}",

year = "2013",

month = feb,

doi = "10.1145/2427023.2427027",

language = "English",

volume = "39",

journal = "ACM Transactions on Mathematical Software",

issn = "0098-3500",

publisher = "Association for Computing Machinery ",

number = "2",

}

TY - JOUR

T1 - Finite element integration on GPUs

AU - Knepley, Matthew G.

AU - Terrel, Andy R.

PY - 2013/2

Y1 - 2013/2

N2 - We present a novel finite element integration method for low-order elements on GPUs. We achieve more than 100GF for element integration on first order discretizations of both the Laplacian and Elasticity operators on an NVIDIA GTX285, which has a nominal single precision peak flop rate of 1 TF/s and bandwidth of 159 GB/s, corresponding to a bandwidth limited peak of 40 GF/s.

AB - We present a novel finite element integration method for low-order elements on GPUs. We achieve more than 100GF for element integration on first order discretizations of both the Laplacian and Elasticity operators on an NVIDIA GTX285, which has a nominal single precision peak flop rate of 1 TF/s and bandwidth of 159 GB/s, corresponding to a bandwidth limited peak of 40 GF/s.

KW - Finite element integration

KW - GPGPUs

UR - https://www.scopus.com/pages/publications/84875154515

U2 - 10.1145/2427023.2427027

DO - 10.1145/2427023.2427027

M3 - Article

AN - SCOPUS:84875154515

SN - 0098-3500

VL - 39

JO - ACM Transactions on Mathematical Software

JF - ACM Transactions on Mathematical Software

IS - 2

M1 - 10

ER -

Finite element integration on GPUs

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this