TY - GEN
T1 - Performance evaluation of cloud-based high performance computing for finite element analysis
AU - Wu, Dazhong
AU - Liu, Xi
AU - Hebert, Steve
AU - Gentzsch, Wolfgang
AU - Terpenny, Janis
N1 - Publisher Copyright:
Copyright © 2015 by ASME.
PY - 2015
Y1 - 2015
N2 - Cloud computing is an innovative computing paradigm that can potentially bridge the gap between increasing computing demands in computer aided engineering (CAE) applications and limited scalability, flexibility, and agility in traditional computing paradigms. In light of the benefits of cloud computing, high performance computing (HPC) in the cloud has the potential to enable users to not only accelerate computationally expensive CAE simulations (e.g., finite element analysis), but also to reduce costs by utilizing on-demand and scalable cloud computing resources. The objective of this research is to evaluate the performance of running a large finite element simulation in a public cloud. Specifically, an experiment is performed to identify individual and interactive effects of several factors (e.g., CPU core count, memory size, solver computational rate, and input/output rate) on run time using statistical methods. Our experimental results have shown that the performance of HPC in the cloud is sufficient for the application of a large finite element analysis, and that run time can be optimized by properly selecting a configuration of CPU, memory, and interconnect.
AB - Cloud computing is an innovative computing paradigm that can potentially bridge the gap between increasing computing demands in computer aided engineering (CAE) applications and limited scalability, flexibility, and agility in traditional computing paradigms. In light of the benefits of cloud computing, high performance computing (HPC) in the cloud has the potential to enable users to not only accelerate computationally expensive CAE simulations (e.g., finite element analysis), but also to reduce costs by utilizing on-demand and scalable cloud computing resources. The objective of this research is to evaluate the performance of running a large finite element simulation in a public cloud. Specifically, an experiment is performed to identify individual and interactive effects of several factors (e.g., CPU core count, memory size, solver computational rate, and input/output rate) on run time using statistical methods. Our experimental results have shown that the performance of HPC in the cloud is sufficient for the application of a large finite element analysis, and that run time can be optimized by properly selecting a configuration of CPU, memory, and interconnect.
UR - http://www.scopus.com/inward/record.url?scp=84979037397&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84979037397&partnerID=8YFLogxK
U2 - 10.1115/DETC2015-46381
DO - 10.1115/DETC2015-46381
M3 - Conference contribution
AN - SCOPUS:84979037397
T3 - Proceedings of the ASME Design Engineering Technical Conference
BT - 35th Computers and Information in Engineering Conference
PB - American Society of Mechanical Engineers (ASME)
T2 - ASME 2015 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2015
Y2 - 2 August 2015 through 5 August 2015
ER -