Computational Technology Resources - CCP

Keywords: parallel computing, clustered network, CORBA, distributed computing, finite element analysis, object oriented technology.

Summary

Network and web technologies have become all pervasive and popular and programming languages embracing object oriented paradigms are very widely used. It is appropriate to consider the scene of scientific finite element computing as a service in a networked environment. While several possibilities exist, the paper addresses the development of such a service in a CORBA environment. Such an environment ensures that the object that the object oriented programming language type and the exact platform of deployment are not important.

A basic object oriented general purpose FE analysis code with a frontal solver has been developed in C++. The element class is projected as the parent class. Three dimensional hexahedron element class, two dimensional six noded LST element class and one dimensional spring element class have been developed as subclasses and successfully tested. The continuum class, which maintains the overall details of the problem, is the starting point of the analysis of the analysis and encapsulates the frontal solver also. The Node class abstraction is used for all issues relate to nodes like nodal coordinates, displacement and stresses.

CORBA has been used as a platform for objects deployment as a network in a client server environment. As mentioned earlier, the server has been developed in C++ while both Java and C++ clients have been used. Different cases viz.

i): Simple client/server model
ii): Multiple concurrent servers with a Java client model have been developed and validated.

The distributed networked environment has been utilized for parallelising the finite element analysis. This has presently been implemented over a cluster of Intel workstations connected through a switch, thereby offering a very economical solution for the solution of large problems. Substructuring and use of super elements have been employed for this parallelisation. Each subserver is responsible for the assembly of elements, computation of super element stiffness and communication with central server. The central server assembles the super element stiffness and evaluates the variables corresponding to the super element nodes. These are communicated back to the subservers which process for the nodes and stresses at the element level. Although the mathematical procedure is well known, implementation in a multi-server multi-client networked environment involves several performance and administrative issues which have been addressed. The advantages of using such an arrangement over a dedicated high performance environment are outlined.

The study also involves the comparison of execution times with two servers. This gives us an idea of the effective utilization of the resources on the network and the ability to solve large problems. Detailed studies have also been conducted with a single client-multi server parallel environment.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £85 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 76 PROCEEDINGS OF THE THIRD INTERNATIONAL CONFERENCE ON ENGINEERING COMPUTATIONAL TECHNOLOGY Edited by: B.H.V. Topping and Z. Bittnar Paper 42 Distributed Framework for Parallel Finite Element Analysis in a CORBA Environment M. Raghunath, C.V. Ramakrishnan, A. Ahlawat and B.P.B. Soma Shekar Department of Applied Mechanics, Indian Institute of Technology, New Delhi, India doi:10.4203/ccp.76.42 purchase the full-text of this paper Full Bibliographic Reference for this paper M. Raghunath, C.V. Ramakrishnan, A. Ahlawat, B.P.B. Soma Shekar, "Distributed Framework for Parallel Finite Element Analysis in a CORBA Environment", in B.H.V. Topping, Z. Bittnar, (Editors), "Proceedings of the Third International Conference on Engineering Computational Technology", Civil-Comp Press, Stirlingshire, UK, Paper 42, 2002. doi:10.4203/ccp.76.42 Keywords: parallel computing, clustered network, CORBA, distributed computing, finite element analysis, object oriented technology. Summary Network and web technologies have become all pervasive and popular and programming languages embracing object oriented paradigms are very widely used. It is appropriate to consider the scene of scientific finite element computing as a service in a networked environment. While several possibilities exist, the paper addresses the development of such a service in a CORBA environment. Such an environment ensures that the object that the object oriented programming language type and the exact platform of deployment are not important. A basic object oriented general purpose FE analysis code with a frontal solver has been developed in C++. The element class is projected as the parent class. Three dimensional hexahedron element class, two dimensional six noded LST element class and one dimensional spring element class have been developed as subclasses and successfully tested. The continuum class, which maintains the overall details of the problem, is the starting point of the analysis of the analysis and encapsulates the frontal solver also. The Node class abstraction is used for all issues relate to nodes like nodal coordinates, displacement and stresses. CORBA has been used as a platform for objects deployment as a network in a client server environment. As mentioned earlier, the server has been developed in C++ while both Java and C++ clients have been used. Different cases viz. i) Simple client/server model ii) Multiple concurrent servers with a Java client model have been developed and validated. The distributed networked environment has been utilized for parallelising the finite element analysis. This has presently been implemented over a cluster of Intel workstations connected through a switch, thereby offering a very economical solution for the solution of large problems. Substructuring and use of super elements have been employed for this parallelisation. Each subserver is responsible for the assembly of elements, computation of super element stiffness and communication with central server. The central server assembles the super element stiffness and evaluates the variables corresponding to the super element nodes. These are communicated back to the subservers which process for the nodes and stresses at the element level. Although the mathematical procedure is well known, implementation in a multi-server multi-client networked environment involves several performance and administrative issues which have been addressed. The advantages of using such an arrangement over a dedicated high performance environment are outlined. The study also involves the comparison of execution times with two servers. This gives us an idea of the effective utilization of the resources on the network and the ability to solve large problems. Detailed studies have also been conducted with a single client-multi server parallel environment. purchase the full-text of this paper (price £20) go to the previous paper go to the next paper return to the table of contents return to the book description purchase this book (price £85 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions