Low cost CPU–GPGPU parallel computing in real-world structural engineering

Abstract Parallel computing has matured from a narrow academic environment, where researchers developed finite element computing tools for performing their studies, to become the basis in modern analysis and design procedures. Several software applications in recent years have made these computing environments accessible to professional engineers, researchers and students outside the computer science research community. These software applications, mainly focused on aerospace, aeronautical, mechanical and naval structural systems, have incorporated the parallel computing component as an additional capability of the finite element software package. On the other hand though, there is not a computational approach for dealing with real-world civil engineering structures such as buildings, bridges or more complex civil engineering structures taking advantage of the high-performance capabilities of contemporary personal computers. Furthermore, graphic processing units (GPUs) have lately been used as high performance co-processors for demanding computational tasks. In modern structural analysis software packages, a hybrid computing model can be implemented, according to which the GPU can be used as a co-processor in order to offload the system's central processing unit (CPU), and therefore increase the overall computational efficiency. SCADA Pro is a structural analysis and design technical software package which implements out-of-core direct and iterative sparse parallel solvers in order to provide high performance computing capabilities in real-world structural engineering. In the present work, real-world structural systems are analysed using SCADA Pro's high performance computing capabilities. The results and computational efficiency are discussed and compared among different solvers, single and multi-threaded CPU and GPU computations.

[1]  M. Papadrakakis,et al.  Accuracy and effectiveness of preconditioned conjugate gradient algorithms for large and ill-conditioned problems , 1993 .

[2]  Mervyn J. Kowalsky,et al.  THE LIMITATIONS AND PERFORMANCES OF DIFFERENT DISPLACEMENT BASED DESIGN METHODS , 2003 .

[3]  Patrick R. Amestoy,et al.  Multifrontal parallel distributed symmetric and unsymmetric solvers , 2000 .

[4]  Robert Park,et al.  REINFORCED CONCRETE MEMBERS WITH CYCLIC LOADING , 1972 .

[5]  P. Tallec.  Domain decomposition methods in computational mechanics , 1994 .

[6]  Manolis Papadrakakis,et al.  An optimized computer implementation of incomplete Cholesky factorization , 1994 .

[7]  Danny C. Sorensen,et al.  Deflation Techniques for an Implicitly Restarted Arnoldi Iteration , 1996, SIAM J. Matrix Anal. Appl..

[8]  M. Hestenes,et al.  Methods of conjugate gradients for solving linear systems , 1952 .

[9]  Nikos D. Lagaros,et al.  A general purpose real-world structural design optimization computing platform , 2014 .

[10]  Charbel Farhat,et al.  Implicit parallel processing in structural mechanics , 1994 .

[11]  Olaf Schenk,et al.  Two-level dynamic scheduling in PARDISO: Improved scalability on shared memory multiprocessing systems , 2002, Parallel Comput..

[12]  Klaus-Jürgen Bathe,et al.  The subspace iteration method - Revisited , 2013 .

[13]  R. Park,et al.  Stress-Strain Behavior of Concrete Confined by Overlapping Hoops at Low and High Strain Rates , 1982 .

[14]  D. Sorensen.  IMPLICITLY RESTARTED ARNOLDI/LANCZOS METHODS FOR LARGE SCALE EIGENVALUE CALCULATIONS , 1996 .

[15]  M. B. Van Gijzen,et al.  Comparison of the deflated preconditioned conjugate gradient method and algebraic multigrid for composite materials , 2012 .

[16]  Michael H. Scott,et al.  Plastic Hinge Integration Methods for Force-Based Beam¿Column Elements , 2006 .

[17]  Manolis Papadrakakis,et al.  The mosaic of high performance domain Decomposition Methods for Structural Mechanics: Formulation, interrelation and numerical efficiency of primal and dual methods , 2003 .

[18]  Barry W. Peyton,et al.  A Supernodal Cholesky Factorization Algorithm for Shared-Memory Multiprocessors , 1991, SIAM J. Sci. Comput..

[19]  Jaehong Lee,et al.  Analysis of tensegrity structures subject to dynamic loading using a Newmark approach , 2015 .

[20]  Nikos D. Lagaros,et al.  An efficient dynamic load balancing algorithm , 2014 .