Two-stage ordering for unsymmetric parallel row-by-row frontal solvers
نویسنده
چکیده
The row-by-row frontal method may be used to solve general large sparse linear systems of equations. By partitioning the matrix into (nearly) independent blocks and applying the frontal method to each block, a coarse-grained parallel frontal algorithm is obtained. The success of this approach depends on preordering the matrix. This can be done in two stages, (1) order the matrix to bordered block diagonal form; (2) order the rows within each block to minimize the size of the frontal matrix. A number of recent papers have considered stage 1. In this paper, an algorithm is proposed for stage 2. For a range of practical examples from chemical process engineering, it is shown that the proposed algorithm substantially reduces the block frontal matrix size and, for sufficiently large problems, this can lead to significant reductions in the factorization times when the row-by-row frontal method is implemented in parallel. © 2001 Elsevier Science Ltd. All rights reserved.
منابع مشابه
A new row ordering strategy for frontal solvers
A new row ordering strategy for frontal solvers. Abstract The frontal method is a variant of Gaussian elimination that has been widely used since the mid 1970s. In the innermost loop of the computation the method exploits dense linear algebra kernels, which are straightforward to vectorize and parallelize. This makes the method attractive for modern computer architectures. However, unless the m...
متن کاملTwo Strategies Based on Meta-Heuristic Algorithms for Parallel Row Ordering Problem (PROP)
Proper arrangement of facility layout is a key issue in management that influences efficiency and the profitability of the manufacturing systems. Parallel Row Ordering Problem (PROP) is a special case of facility layout problem and consists of looking for the best location of n facilities while similar facilities (facilities which has some characteristics in common) should be arranged in a row ...
متن کاملFrontal Solvers for Process Engineering: Local Row Ordering Strategies
The solution of chemical process simulation and optimization problems on today's high performance supercomputers requires algorithms that can take advantage of vector and parallel processing when solving the large, sparse matrices that arise. The frontal method can be highly e cient in this context due to its ability to make use of vectorizable dense matrix kernels on a relatively small frontal...
متن کاملRow Ordering for Frontal Solvers in Chemical Process Engineering
The solution of chemical process engineering problems often requires the repeated solution of large sparse linear systems of equations that have a highly asymmetric structure. The frontal method can be very e cient for solving such systems on modern computer architectures because, in the innermost loop of the computation, the method exploits dense linear algebra kernels, which are straightforwa...
متن کاملA Large-Grain Parallel Sparse System Solver
The eeciency of solving sparse linear systems on parallel processors and more complex multicluster architectures such as Cedar is greatly enhanced if relatively large grain computational tasks can be assigned to each cluster or processor. The ordering of a system into a bordered block upper triangular form facilitates a reasonable large-grain partitioning. A new algorithm which produces this fo...
متن کامل