This paper discusses fast parallel algorithms for evaluating several centrality indices frequently used in complex network analysis. These algorithms have been optimized to exploit properties typically observed in rea...
详细信息
ISBN:
(纸本)0769526365
This paper discusses fast parallel algorithms for evaluating several centrality indices frequently used in complex network analysis. These algorithms have been optimized to exploit properties typically observed in real-world large scale networks, such as the low average distance, high local density, and heavy-tailed power law degree distributions. We test our implementations on real datasets such as the web graph, protein-interaction networks, movie-actor and citation networks, and report impressive parallel performance for evaluation of the computationally intensive centrality metrics (betweenness and closeness centrality) on high-end shared memory symmetric multiprocessor and multithreaded architectures. To our knowledge, these are the first parallel implementations of these widely-used social network analysis metrics. We demonstrate that it is possible to rigorously analyze networks three orders of magnitude larger than instances that can be handled by existing network analysis (SNA) software packages. For instance, we compute the exact betweenness centrality value for each vertex in a large US patent citation network (3 million patents, 16 million citations) in 42 minutes on 16 processors, utilizing 20GB RAM of the IBM p5 570. Current SNA packages on the other hand cannot handle graphs with more than hundred thousand edges.
Neural algorithmic reasoners are parallel processors. Teaching them sequential algorithms contradicts this nature, rendering a significant share of their computations redundant. parallel algorithms however may exploit...
详细信息
Neural algorithmic reasoners are parallel processors. Teaching them sequential algorithms contradicts this nature, rendering a significant share of their computations redundant. parallel algorithms however may exploit their full computational power, therefore requiring fewer layers to be executed. This drastically reduces training times, as we observe when comparing parallel implementations of searching, sorting and finding strongly connected components to their sequential counterparts on the CLRS framework. Additionally, parallel versions achieve (often strongly) superior predictive performance.
In this paper we have developed algorithms to solve macroeconometric models with forward-looking variables based on Newton method for nonlinear systems of equations. The most difficult step for Newton methods represen...
详细信息
In this paper we have developed algorithms to solve macroeconometric models with forward-looking variables based on Newton method for nonlinear systems of equations. The most difficult step for Newton methods represents the resolution of a large linear system for each iteration. Thus, we compare the performances resulted by solving this linear system using two iterative methods and the direct method. We've also described an implementation of the parallel versions of such algorithms using a software package. Our experiments confirm that the iterative methods have a low computational complexity and storage requirements, but the parallel versions of direct methods (C) 2010 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of the Guest Editor.
The dynamical properties of many natural phenomena can be related to their support fractal dimension. A relevant example is the connection between flood peaks produced in a river basin, as observed in flood hydrograph...
详细信息
ISBN:
(纸本)9783030390815
The dynamical properties of many natural phenomena can be related to their support fractal dimension. A relevant example is the connection between flood peaks produced in a river basin, as observed in flood hydrographs, and the multi-fractal spectrum of the river itself, according to the Multifractal Instantaneous Unit Hydrograph (MIUH) theory. Typically, the multifractal analysis of river networks is carried out by sampling large collections of points belonging to the river basin and analyzing the fractal dimensions and the Lipschitz-Holder exponents of singularities through numerical procedures which involve different degrees of accuracy in the assessment of such quantities through different methods (box-counting techniques, the generalized correlation integral method by Pawelzik and Schuster (1987), the fixed-mass algorithms by Badii and Politi (1985), being some relevant examples). However, the higher accuracy in the determination of the fractal dimensions requires considerably higher computational times. For this reason, we recently developed a parallel version of some of the cited multifractal methods described above by using the MPI parallel library, by reaching almost optimal speed-ups in the computations. This will supply a tool for the assessment of the fractal dimensions of river networks (as well as of several other natural phenomena whose embedding dimension is 2 or 3) on massively parallel clusters or multi-core workstations.
Efficient parallel algorithms for computing all possible subset regression models are proposed. The algorithms are based on the dropping columns method that generates a regression tree. The properties of the tree are ...
详细信息
Efficient parallel algorithms for computing all possible subset regression models are proposed. The algorithms are based on the dropping columns method that generates a regression tree. The properties of the tree are exploited in order to provide an efficient load balancing which results in no inter-processor communication. Theoretical measures of complexity suggest linear speedup. The parallel algorithms are extended to deal with the general linear and seemingly unrelated regression models. The case where new variables are added to the regression model is also considered. Experimental results on a shared memory machine are presented and analyzed. (C) 2003 Elsevier Science B.V. All rights reserved.
We present MPI-based parallel algorithms for counting triangles and computing clustering coefficients in massive networks. Counting triangles is important in the analysis of various networks, e.g., social, biological,...
详细信息
We discuss and compare two approaches for model reduction of large-scale unstable systems on parallel computers. The first method proceeds by computing the additive decomposition of the transfer function via a block d...
详细信息
ISBN:
(纸本)0780395670
We discuss and compare two approaches for model reduction of large-scale unstable systems on parallel computers. The first method proceeds by computing the additive decomposition of the transfer function via a block diagonalization, followed by a reduction of the stable part of the system using techniques based on state-space truncation. The second method employs a representation of the controllability and observability Gramians of an unstable systems in terms of the Gramians of the stabilized system where the particular stabilization is obtained via the solution of dual algebraic Bernoulli equations. Based on these Gramians, balanced truncation is then applied in the usual manner. All core computational steps in these methods can be efficiently solved on parallel computers by means of diverse variants of the Newton iteration for the sign function. Numerical experiments on a cluster of Intel Xeon processors show the numerical and parallel performances of these methods.
The ability to model the temporal dimension is essential to many applications. Furthermore, the rate of increase in database size and response time requirements has outpaced advancements in processor and mass storage ...
详细信息
ISBN:
(纸本)0769500714
The ability to model the temporal dimension is essential to many applications. Furthermore, the rate of increase in database size and response time requirements has outpaced advancements in processor and mass storage technology, leading to the need for parallel temporal database management systems. In this paper toe introduce a variety of parallel temporal aggregation algorithms for a shared-nothing architecture based on the sequential Aggregation Tree algorithm. Via an empirical study, we found that the number of processing nodes, the partitioning of the data, the placement of results, and the degree of data reduction effected by the aggregation impacted the performance of the algorithms. For distributed results placement, we discovered that Time Division Merge was the obvious choice. For centralized results and high data reduction, Pairwise Merge was preferred regardless of the number of processing nodes, but for low data reduction, it only performed well up to 32 nodes. This led us to a centralized variant of Time Division Merge which was best for larger configurations having low data reduction.
Given an n-vertex simple polygon we address the following problems: (i) find the shortest path between two points s and d inside P, and (ii) compute the shortestpath tree between a single point s and each vertex of P ...
详细信息
We present parallel algorithms for computing cycle orders and cycle perimeters in relative neighborhood graphs. This parallel algorithm has wide-ranging applications from microscopic to macroscopic domains, e.g., in h...
详细信息
ISBN:
(纸本)9781538610428
We present parallel algorithms for computing cycle orders and cycle perimeters in relative neighborhood graphs. This parallel algorithm has wide-ranging applications from microscopic to macroscopic domains, e.g., in histopathological image analysis and wireless network routing. Our algorithm consists of the following steps (sub-algorithms): (1) Uniform partitioning of the graph vertices across processes, (2) parallel Delaunay triangulation and (3) parallel computation of the relative neighborhood graph and the cycle orders and perimeters. We evaluated our algorithm on a large dataset with 6.5 Million points and demonstrate excellent fixed-size scalability. We also demonstrate excellent isogranular scalability up to 131K processes. Our largest run was on a dataset with 13 billion points on 131K processes on ORNL's Cray XK7 "Titan" supercomputer.
暂无评论