Data Communication and Parallel Computing on Twisted Hypercubes
نویسنده
چکیده
Massively parallel distributed-memory architectures are receiving increasing attention to meet the increasing demand on processing power. Many topologies have been proposed for interconnecting the processors of distributed computing systems. The hypercube topology has drawn considerable attention due to many of its attractive properties. The appealing properties of the hypercube topology such as vertex and edge symmetry, recursive structure, logarithmic diameter, maximally fault-tolerance, simple routing and broadcasting, and the ability to simulate other interconnection networks with minimum overhead have made it an excellent candidate for many parallel processing applications. Many variations of the hypercube topology have been reported in the literature, mainly to add to the computational power of the hypercube. One of the attractive versions of the hypercube that was introduced to enhance the performance is the twisted hypercube. A twisted hypercube has the same structural complexities of the hypercube. It preserves the attractive properties of the hypercube and improves on the communication time by reducing the diameter by a factor of two. This paper presents the basic communication and some of the basic operations usually needed in parallel computing on the twisted hypercube interconnection network.
منابع مشابه
Improved Probabilistic Routing on Generalized Hypercubes
p. 1 Efficient Data Communication in Incomplete Hypercubes p. 13 Efficient Communication in the Folded Petersen Interconnection Networks p. 25 Compiling Rewriting onto SIMD and MIMD/SIMD Machines p. 37 A Compilation Technique for Varying Communication Cost NUMA Architectures p. 49 A Data Partitioning Algorithm for Distributed Memory Compilation p. 61 Towards a High Precision Massively Parallel ...
متن کاملHamiltonian Connectivity of Twisted Hypercube-Like Networks under the Large Fault Model
Twisted hypercube-like networks (THLNs) are an important class of interconnection networks for parallel computing systems, which include most popular variants of the hypercubes, such as crossed cubes, Möbius cubes, twisted cubes and locally twisted cubes. This paper deals with the fault-tolerant hamiltonian connectivity of THLNs under the large fault model. Let G be an n-dimensional THLN and F ...
متن کاملData mapping of linear programming on fixed-size hypercubes
Although many solution methods are available for the linear programming problem, the simplex method is undoubtedly the most widely used one for its simplicity. In this paper, we propose an implementation of the simplex method on freed-size hypercubes. A partitioning technique and a mapping technique are presented to fit largo~sL~ problem instances into relatively small-size hypercubes. Two case...
متن کاملA Simple Parallel Algorithm for Constructing Independent Spanning Trees on Twisted Cubes∗
In 1989, Zehavi and Itai [46] proposed the following conjecture: a k-connected graph G must possess k independent spanning trees (ISTs for short) with an arbitrary node as the common root. An n-dimensional twisted cube, denoted by TQn, is a variation of hypercubes with connectivity n to achieving some improvements of structure properties. Recently, Yang [42] proposed an algorithm for constructi...
متن کاملA Formal Study of the Mcube Interconnection Network
The Mcube network has been proposed in [11] as a highly recursive and symmetrical interconnection network based on twisted links [1]. The Mcube topology has been developed and defined in terms of the structural constraints between components to enforce structural symmetry instead of specifying edges between node pairs. Mcubes have almost half the diameter of a comparable hypercube, a lower aver...
متن کامل