Parallel FDTD Computations Based on Macro Data Flow Graphs Transformed by Genetic Algorithms
نویسندگان
چکیده
The Finite Di erence Time Domain (FDTD) method belongs to numerical applications that can be described by irregular data patterns or use irregular data structures. This method is based on an iterative algorithm, which is used to simulate high frequency electromagnetic wave propagation. Such simulation is performed by solving Maxwell equations transformed into their di erential form. In order to perform such a simulation in a multiprocessor system, the whole computational area must be divided into subareas assigned to particular processors. In order to balance computational load, size of all computational subareas must be equalized. To minimize communication volume, length of the border between each two adjacent subareas must be minimal.
منابع مشابه
Modeling and scheduling no-idle hybrid flow shop problems
Although several papers have studied no-idle scheduling problems, they all focus on flow shops, assuming one processor at each working stage. But, companies commonly extend to hybrid flow shops by duplicating machines in parallel in stages. This paper considers the problem of scheduling no-idle hybrid flow shops. A mixed integer linear programming model is first developed to mathematically form...
متن کاملA New ILP Model for Identical Parallel-Machine Scheduling with Family Setup Times Minimizing the Total Weighted Flow Time by a Genetic Algorithm
This paper presents a novel, integer-linear programming (ILP) model for an identical parallel-machine scheduling problem with family setup times that minimizes the total weighted flow time (TWFT). Some researchers have addressed parallel-machine scheduling problems in the literature over the last three decades. However, the existing studies have been limited to the research of independent jobs,...
متن کاملOptimization of Parallel FDTD Computations Based on Program Macro Data Flow Graph Transformations
This chapter concerns numerical problems that are solved by parallel regular computations performed in rectangular meshes that span over irregular computational areas. Such parallel problems are more difficult to be optimized than problems concerning regular areas since the problem cannot be solved by a simple geometrical decomposition of the computational area. Usually, a kind of step-by-step ...
متن کاملRuntime Data Flow Scheduling of Matrix Computations
We investigate the scheduling of matrix computations expressed as directed acyclic graphs for shared-memory parallelism. Because of the data granularity in this problem domain, even slight variations in load balance or data locality can greatly affect performance. Well-known scheduling algorithms such as work stealing have proven time and space bounds, but these bounds do not provide a discerna...
متن کاملParleda: a Library for Parallel Processing in Computational Geometry Applications
ParLeda is a software library that provides the basic primitives needed for parallel implementation of computational geometry applications. It can also be used in implementing a parallel application that uses geometric data structures. The parallel model that we use is based on a new heterogeneous parallel model named HBSP, which is based on BSP and is introduced here. ParLeda uses two main lib...
متن کامل