Haishan Lu, Shuguang Gong^*, Jianping Zhang, Guilan Xie, Shuohui Yin

CMES-Computer Modeling in Engineering & Sciences, Vol.128, No.3, pp. 1151-1178, 2021, DOI:10.32604/cmes.2021.016165 - 11 August 2021

Abstract We proposed an improved graphics processing unit (GPU) acceleration approach for three-dimensional structural topology optimization using the element-free Galerkin (EFG) method. This method can effectively eliminate the race condition under parallelization. We established a structural topology optimization model by combining the EFG method and the solid isotropic microstructures with penalization model. We explored the GPU parallel algorithm of assembling stiffness matrix, solving discrete equation, analyzing sensitivity, and updating design variables in detail. We also proposed a node pair-wise method for assembling the stiffness matrix and a node-wise method for sensitivity analysis to eliminate race conditions More >

G.A. Gravvanis¹, K.M. Giannoutakis²

CMES-Computer Modeling in Engineering & Sciences, Vol.71, No.4, pp. 305-330, 2011, DOI:10.3970/cmes.2011.071.305

Abstract In this paper we present parallel explicit approximate inverse matrix techniques for solving sparse linear systems on shared memory systems, which are derived using the finite element method for biharmonic equations in three space variables. Our approach for solving such equations is by considering the biharmonic equation as a coupled equation approach (pair of Poisson equation), using a FE approximation scheme, yielding an inner-outer iteration method. Additionally, parallel approximate inverse matrix algorithms are introduced for the efficient solution of sparse linear systems, based on an anti-diagonal computational approach that eliminates the data dependencies. Parallel explicit More >

S.H. Ju¹

CMES-Computer Modeling in Engineering & Sciences, Vol.64, No.1, pp. 91-109, 2010, DOI:10.3970/cmes.2010.064.091

Abstract This study develops an OpenMP scheme to parallel the preconditioned conjugate gradient methods (PCG) in shared memory computers. The proposed method is simple and systematic, so a minor change in traditional PCG methods may produce effective parallelism. At first, the global stiffness matrix is re-numbered in order to produce a parallel three-line form matrix, and a subroutine only needs to be called once in the finite element analysis. Several basic OpenMP commands are then added into the traditional incomplete Cholesky factorization (ILU) and symmetric successive over-relaxation (SSOR) codes to make the procedures of matrix multiplication, More >

T.B Jönsthövel¹, M.B. van Gijzen², C.Vuik², C. Kasbergen¹, A. Scarpas¹

CMES-Computer Modeling in Engineering & Sciences, Vol.47, No.2, pp. 97-118, 2009, DOI:10.3970/cmes.2009.047.097

Abstract The introduction of computed x-ray tomography allows for the construction of high quality, material-per-element based 3D meshes in the field of structural mechanics. The use of these meshes enables a shift from meso to micro scale analysis of composite materials like cement concrete, rocks and asphalt concrete. Unfortunately, because of the extremely long execution time, memory and storage space demands, the majority of commercially available finite element packages are not capable of handling efficiently the most computationally demanding operation of the finite element solution process, that is, the inversion of the structural stiffness matrix. To More >

G.A. Gravvanis, K.M. Giannoutakis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.32, No.1, pp. 35-44, 2008, DOI:10.3970/cmes.2008.032.035

Abstract A new parallel normalized optimized approximate inverse algorithm, based on the concept of the ``fish bone'' computational approach with cyclic distribution of the processors satisfying an antidiagonal data dependency, for computing classes of explicit approximate inverses, is introduced for symmetric multiprocessor systems. The parallel normalized explicit approximate inverses are used in conjunction with parallel normalized explicit preconditioned conjugate gradient square schemes, for the efficient solution of finite element sparse linear systems. The parallel design and implementation issues of the new proposed algorithms are discussed and the parallel performance is presented, using OpenMP. More >

G.A. Gravvanis¹, K.M. Giannoutakis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.16, No.2, pp. 69-82, 2006, DOI:10.3970/cmes.2006.016.069

Abstract A new class of normalized explicit optimized approximate inverse finite element matrix techniques, based on normalized finite element approximate factorization procedures, for solving sparse linear systems resulting from the finite element discretization of partial differential equations in three space variables are introduced. A new parallel normalized explicit preconditioned conjugate gradient square method in conjunction with normalized approximate inverse finite element matrix techniques for solving efficiently sparse finite element linear systems on distributed memory systems is also presented along with theoretical estimates on speedups and efficiency. The performance on a distributed memory machine, using Message Passing More >