Bandwidth and Pro le Reduction of Sparse Matrices :
نویسندگان
چکیده
The problem of sparse matrix bandwidth reduction is addressed and solved with two approaches suitable for both symmetric and non-symmetric matrices. The former is a constructive method, the latter is an application of a metaheuristic scheme. This heuristic algorithm has been implemented on parallel machines. Both approaches are compared with previous methods in the literature (suited only for symmetric cases). Benchmarks on real industrial cases, show the superior eeectiveness and eeciency of the proposed methods. Some results illustrating the parallel performance of the parallel heuristic are also reported. The extension of the proposed algorithms to the proole minimization problem is presented.
منابع مشابه
Simulated Annealing for Pro le and Fill Reduction of Sparse Matrices
1 2 Summary Simulated annealing can minimize both proole and ll of sparse matrices. We applied these techniques to a number of sparse matrices from the Harwell-Boeing Sparse Matrix Collection. We were able to reduce proole typically to about 80% of that attained by conventional proole minimization techniques (and sometimes much lower), but ll reduction was less successful (85% at best). We pres...
متن کاملBandwidth and Wavefront Reduction for Static Variable Ordering in Symbolic Model Checking
We demonstrate the applicability of bandwidth and wavefront reduction algorithms to static variable ordering. In symbolic model checking event locality plays a major role in time and memory usage. For example, in Petri nets event locality can be captured by dependency matrices, where nonzero entries indicate whether a transition modifies a place. The quality of event locality has been expressed...
متن کاملSparse Matrix-Vector Multiplication on FPGAs
Floating-point Sparse Matrix-Vector Multiplication (SpMXV) is a key computational kernel in scientic and engineering applications. The poor data locality of sparse matrices signicantly reduces the performance of SpMXV on general-purpose processors, which rely heavily on the cache hierarchy to achieve high performance. The abundant hardware resources on current FPGAs provide new opportunities to...
متن کاملAverage Bandwidth Reduction in Sparse Matrices Using Hybrid Heuristics
This paper proposes a hybrid heuristic aiming the reduction of the average bandwidth of sparse matrices. The proposed heuristic combines a greedy selection of rows and columns to be interchanged with an approach based on a genetic-inspired crossover. Several ideas of limiting the search tree are also investigated. Preliminary numerical experiments illustrate the fact that the proposed heuristic...
متن کاملOrdering for Bandwidth and Profile Minimization Problems via Charged System Search Algorithm
In this paper the recently developed meta-heuristic optimization method, known as charged system search (CSS), is utilized for optimum nodal ordering to minimize bandwidth and profile of sparse matrices. The CSS is an optimization algorithm, which is based on the governing laws of Coulomb and Gauss from electrostatics and the Newtonian mechanics of motion. The bandwidth and profile of some grap...
متن کامل