Parallel Algorithms on Three - Levelhierarchical Cubic
نویسنده
چکیده
In this paper, we propose a novel interconnection network called three-level hierarchical cubic network (3-HCN), which extends the construction of hierarchical cubic networks (HCN) and ooers many additional advantages. The 3-HCN (with diameter links) has diameter smaller than that of an n-cube of the same size and uses nodes of degree 1 3 n + 2. We develop eecient and elegant algorithms for routing, broadcasting, semigroup computation, and matrix-matrix multiplication on 3-HCN. In particular, 3 p N 3 p N matrix-matrix multiplication can be performed in 1 3 log 2 N + 2 time, which is approximately one fth of that required for the DNS algorithm on the hypercube and HCN. We also develop algorithms for emulating a hyper-cube of the same size in O(1) time on 3-HCN. We present variants of 3-HCN that use diierent modules, such as tree, mesh, and buslet, to t the needs of various applications. We conclude that the 3-HCN and its variants are suitable for constructing high-performance special-purpose systems (e.g., for numerical computations) as well as cost-eeective general-purpose parallel architectures.
منابع مشابه
Parallel Matrix Multiplication Algorithms on Hypercube Multiprocessors 1
In this paper, we present three parallel algorithms for matrix multiplication. The rst one, which employs pipelining techniques on a mesh grid, uses only one copy of data matrices. The second one uses multiple copies of data matrices also on a mesh grid. Although data communication operations of the second algorithm are reduced, the requirement of local data memory for each processing element i...
متن کاملA New Parallel Matrix Multiplication Method Adapted on Fibonacci Hypercube Structure
The objective of this study was to develop a new optimal parallel algorithm for matrix multiplication which could run on a Fibonacci Hypercube structure. Most of the popular algorithms for parallel matrix multiplication can not run on Fibonacci Hypercube structure, therefore giving a method that can be run on all structures especially Fibonacci Hypercube structure is necessary for parallel matr...
متن کاملSub - cubic Cost Algorithms for the All Pairs
In this paper we give three sub-cubic cost algorithms for the all pairs shortest distance (APSD) and path (APSP) problems. The rst is a parallel algorithm that solves the APSD problem for a directed graph with unit edge costs in O(log 2 n) time with O(n p logn) processors where = 2:688 on an EREW-PRAM. The second parallel algorithm solves the APSP, and consequently APSD, problem for a directed ...
متن کاملParallel Implementation of Particle Swarm Optimization Variants Using Graphics Processing Unit Platform
There are different variants of Particle Swarm Optimization (PSO) algorithm such as Adaptive Particle Swarm Optimization (APSO) and Particle Swarm Optimization with an Aging Leader and Challengers (ALC-PSO). These algorithms improve the performance of PSO in terms of finding the best solution and accelerating the convergence speed. However, these algorithms are computationally intensive. The go...
متن کاملA Parallel Framework for Multipoint Spiral Search in ab Initio Protein Structure Prediction
Protein structure prediction is computationally a very challenging problem. A large number of existing search algorithms attempt to solve the problem by exploring possible structures and finding the one with the minimum free energy. However, these algorithms perform poorly on large sized proteins due to an astronomically wide search space. In this paper, we present a multipoint spiral search fr...
متن کامل