Totally Parallel Multilevel Algorithms
نویسنده
چکیده
منابع مشابه
Parallel multilevel algorithms for hypergraph partitioning
In this paper, we present parallel multilevel algorithms for the hypergraph partitioning problem. In particular, we describe schemes for parallel coarsening, parallel greedy k-way refinement and parallel multi-phase refinement. Using an asymptotic theoretical performance model, we derive the isoefficiency function for our algorithms and hence show that they are technically scalable when the max...
متن کاملParallel Multilevel Block ILU Preconditioning Techniques for Large Sparse Linear Systems
We present a class of parallel preconditioning strategies built on a multilevel block incomplete LU (ILU) factorization technique to solve large sparse linear systems on distributed memory parallel computers. The preconditioners are constructed by using the concept of block independent sets. Two algorithms for constructing block independent sets of a distributed sparse matrix are proposed. We c...
متن کاملSerial and Parallel Multilevel Graph Partitioning Using Fixed Centers
We present new serial and parallel algorithms for multilevel graph partitioning. Our algorithm has coarsening, partitioning and uncoarsening phases like other multilevel partitioning methods. However, we choose fixed nodes which are at least a specified distance away from each other and coarsen them with their neighbor nodes in the coarsening phase using various heuristics. Using this algorithm...
متن کاملReduction the Number of Power Electronic Devices of a Cascaded Multilevel Inverter Based on New General Topology
In this paper, a new cascaded multilevel inverter by capability of increasing the number of output voltage levels with reduced number of power switches is proposed. The proposed topology consists of series connection of a number of proposed basic multilevel units. In order to generate all voltage levels at the output, five different algorithms are proposed to determine the magnitude of DC volta...
متن کاملConstructive Interference in Parallel Algorithms *
Parallel algorithms are developed in the setting of iterative multilevel methods. The constituent parts of the algorithms are dependent rather than independent as in conventional parallel algorithms. We develop cases and conditions wherein this dependence generates a constructive interference in the computation. The resulting parallel algorithms can then be more efficient than serial counterparts.
متن کامل