检索结果-内蒙古大学图书馆

parallel Regularized Multiple-criteria Linear Programming

Procedia Computer Science 2014年 31卷 58-65页

作者： Zhiquan Qi Vassil Alexandrov Yong Shi Yingjie Tian Research Center on Fictitious Economy & Data Science Chinese Academy of Sciences Beijing 100190 China ICREA and Barcelona Supercomputing Center C/Jordi Girona 29 Edifici Nexus II E-08034 Barcelona Spain

In this paper, we proposed a new parallel algorithm: parallel Regularized Multiple-Criteria Linear Programming (PRMCLP) to overcome the computing and storage requirements increased rapidly with the number of training samples. Firstly, we convert RMCLP model into a unconstrained optimization problem, and then split it into several parts, and each part is computed by a single processor. After that, we analyze each part's result for next cycle going. By doing this, we are be able to obtain the final optimization solution of the whole classification problem. All experiments in public datasets show that our method greatly increases the training speed of RMCLP in the help of multiple processors.

关键词： PRMCLP parallel algorithm data mining

来源：评论

学校读者我要写书评

暂无评论

NUMERICAL SIMULATION OF UNSTEADY WIND-INDUCED CONDUCTOR OSCILLATIONS 11

NUMERICAL SIMULATION OF UNSTEADY WIND-INDUCED CONDUCTOR OSCI...

引用

11th World Congress on Computational Mechanics (WCCM) / 5th European Conference on Computational Mechanics (ECCM) / 6th European Conference on Computational Fluid Dynamics (ECFD)

作者： Ivanova, Olga A. Bauman Moscow State Tech Univ Dept Appl Math Moscow 105005 Russia

ISBN: (纸本)9788494284472

The mathematical model of aeroelastic motion of the transmission line conductor is considered. parallel program for the conductor motion simulation is developed. The unsteady aerodynamic loads on the conductor cross-sections were calculated using viscous vortex domains method. Some test computations were performed. The results of the computations were compared to known experimental and numerical data.

关键词： Mathematical Modeling Transmission Line Galloping Vortex Element Viscous Vortex Domains Method parallel algorithm MPI Technology

来源：评论

学校读者我要写书评

暂无评论

A! - A Cooperative Heuristic Search algorithm 7

A! - A Cooperative Heuristic Search Algorithm

引用

7th European Starting AI Researcher Symposium (STAIRS)

作者： Halme, Antti Aalto Univ FI-00076 Aalto Finland

ISBN: (纸本)9781614994213;9781614994206

We propose a new parallel search algorithm - A! - based on cooperating A* search agents, concurrency and a secondary tiebreaking heuristic. The search agents in A! share information asynchronously and trade some of their independence for additional search focus and a more global view of the search task. A! is inherently nondeterministic due to the implicit randomness of instruction scheduling, but given a consistent primary heuristic, it still finds optimal solutions for the single- source shortest path problem (SSSP). A! combines into a single cooperative search algorithm the breadth available in parallel execution and the depth- first orientation of both locally and globally informed search. We experimentally show that A! outperforms both vanilla A* and an explicitly randomized, noncooperative parallel A* variant. We present an empirical study on cooperation benefits and scalability in the classic 15- puzzle context. The results imply that cooperation and concurrency can successfully be harnessed in algorithm design, inviting further inquiry into algorithms of this kind.

关键词： A heuristic search parallel algorithm cooperation nondeterminism

来源：评论

学校读者我要写书评

暂无评论

Load Balancing N-Body Simulations with Highly Non-Uniform Density 14

Load Balancing N-Body Simulations with Highly Non-Uniform De...

引用

28th ACM International Conference on Supercomputing (ICS)

作者： Pearce, Olga Gamblin, Todd de Supinski, Bronis R. Arsenlis, Tom Amato, Nancy M. Texas A&M Univ Dept Comp Sci & Engn College Stn TX 77843 USA Lawrence Livermore Natl Lab Ctr Appl Sci Comp Livermore CA USA

ISBN: (纸本)9781450326421

N-body methods simulate the evolution of systems of particles (or bodies). They are critical for scientific research in fields as diverse as molecular dynamics, astrophysics, and material science. Most load balancing techniques for N-body methods use particle count to approximate computational work. This approximation is inaccurate, especially for systems with high density variation, because work in an N-body simulation is proportional to the particle density, not the particle count. In this paper, we demonstrate that existing techniques do not perform well at scale when particle density is highly non-uniform, and we propose a load balance technique that efficiently assigns load in terms of interactions instead of particles. We use adaptive sampling to create an even work distribution more amenable to partitioning, and to reduce partitioning overhead. We implement and evaluate our approach on a Barnes-Hut algorithm and a large-scale dislocation dynamics application, ParaDiS. Our method achieves up to 26% improvement in overall performance of Barnes-Hut and 18% in ParaDiS.

关键词： load balance parallel algorithm performance simulation

来源：评论

学校读者我要写书评

暂无评论

Voxel based parallel post processor for void nucleation and growth analysis of atomistic simulations of material fracture

引用

JOURNAL OF MOLECULAR GRAPHICS & MODELLING 2014年 50卷 134-141页

作者： Hemani, H. Warrier, M. Sakthivel, N. Chaturvedi, S. Bhabha Atom Res Ctr Visakhapatnam 530012 Andhra Pradesh India

Molecular dynamics (MD) simulations are used in the study of void nucleation and growth in crystals that are subjected to tensile deformation. These simulations are run for typically several hundred thousand time steps depending on the problem. We output the atom positions at a required frequency for post processing to determine the void nucleation, growth and coalescence due to tensile deformation. The simulation volume is broken up into voxels of size equal to the unit cell size of crystal. In this paper, we present the algorithm to identify the empty unit cells (voids), their connections (void size) and dynamic changes (growth and coalescence of voids) for MD simulations of large atomic systems (multi-million atoms). We discuss the parallel algorithms that were implemented and discuss their relative applicability in terms of their speedup and scalability. We also present the results on scalability of our algorithm when it is incorporated into MD software LAMMPS. (C) 2014 Elsevier Inc. All rights reserved.

关键词： Voxel Void nucleation and growth parallel algorithm Molecular dynamics Fracture

来源：评论

学校读者我要写书评

暂无评论

A Decomposition algorithm for Mean-Variance Economic Model Predictive Control of Stochastic Linear Systems

A Decomposition Algorithm for Mean-Variance Economic Model P...

引用

IEEE International Symposium on Intelligent Control (ISIC)

作者： Sokoler, Leo Emil Dammann, Bernd Madsen, Henrik Jorgensen, John Bagterp Tech Univ Denmark Dept Appl Math & Comp Sci DK-2800 Lyngby Denmark DONG Energy DK-2820 Gentofte Denmark

ISBN: (纸本)9781479974061

This paper presents a decomposition algorithm for solving the optimal control problem (OCP) that arises in Mean-Variance Economic Model Predictive Control of stochastic linear systems. The algorithm applies the alternating direction method of multipliers to a reformulation of the OCP that decomposes into small independent subproblems. We test the decomposition algorithm using a simple power management case study, in which the OCP is formulated as a convex quadratic program. Simulations show that the decomposition algorithm scales linearly in the number of uncertainty scenarios. Moreover, a parallel implementation of the algorithm is several orders of magnitude faster than state-of-the-art convex quadratic programming algorithms, provided that the number of uncertainty scenarios is large.

关键词： convex programming linear systems optimal control parallel algorithms predictive control quadratic programming stochastic systems OCP reformulation convex quadratic programming algorithms decomposition algorithm mean-variance economic model predictive control multipliers alternating direction method optimal control problem parallel algorithm power management stochastic linear systems subproblems Linear programming Linear systems Optimization Prediction algorithms Trajectory Uncertainty Vectors Linear system parallel algorithms Predictive control quadratic programming Optimal control Prediction algorithms convex programming stochastic systems linear programming trajectory Open circuit potential decomposition Economic Models

来源：评论

学校读者我要写书评

暂无评论

Preliminary Application of Software Package GPPS to Spare Linear Systems from Meso Scale Simulation of Concrete Specimen

Preliminary Application of Software Package GPPS to Spare Li...

引用

4th International Conference on Civil Engineering, Architecture and Building Materials (CEABM)

作者： Wu, Jianping Ma, Huaifa Zhao, Jun Wang, Shuchang Natl Univ Def Technol Sch Comp Sci Changsha 410073 Hunan Peoples R China

ISBN: (纸本)9783038351658

The general parallel preconditioning subspace iteration software GPPS is developed based on the integration of various preconditioning methods, parallelization techniques, and subspace iterations, to solve general sparse linear systems. The software has several characteristics, including high computational performance, good readability, simple interface, and excellent scalability. In this paper, the functions of the software are outlined, and then it is used to solve the sparse linear systems from meso-scale simulation of concrete specimens. The numerical experiments show that GPPS is clearly superior to the software package AZTEC, which is used in the simulation up to now, and in addition, ICT and the parallelization technique based on factors combination outperform other preconditioning techniques integrated in the software.

关键词： parallel algorithm preconditioner subspace iteration sparse linear system meso-scale simulation of concrete

来源：评论

学校读者我要写书评

暂无评论

Training Large Scale Deep Neural Networks on the Intel Xeon Phi Many-core Coprocessor 28

Training Large Scale Deep Neural Networks on the Intel Xeon ...

引用

28th IEEE International parallel & Distributed Processing Symposium Workshops (IPDPSW)

作者： Jin, Lei Wang, Zhaokang Gu, Rong Yuan, Chunfeng Huang, Yihua Nanjing Univ Dept Comp Sci & Technol Natl Key Lab Novel Software Technol Nanjing 210023 Jiangsu Peoples R China

ISBN: (纸本)9781479941162

As a new area of machine learning research, the deep learning algorithm has attracted a lot of attention from the research community. It may bring human beings to a higher cognitive level of data. Its unsupervised pre-training step allows us to find high-dimensional representations or abstract features which work much better than the principal component analysis (PCA) method. However, it will face problems when being applied to deal with large scale data due to its intensive computation from many levels of training process against large scale data. The sequential deep learning algorithms usually can not finish the computation in an acceptable time. In this paper, we propose a many-core algorithm which is based on a parallel method and is used in the Intel Xeon Phi many-core systems to speed up the unsupervised training process of Sparse Autoencoder and Restricted Boltzmann Machine (RBM). Using the sequential training algorithm as a baseline to compare, we adopted several optimization methods to parallelize the algorithm. The experimental results show that our fully-optimized algorithm gains more than 300-fold speedup on parallelized Sparse Autoencoder compared with the original sequential algorithm on the Intel Xeon Phi coprocessor. Also, we ran the fully-optimized code on both the Intel Xeon Phi coprocessor and an expensive Intel Xeon CPU. Our method on the Intel Xeon Phi coprocessor is 7 to 10 times faster than the Intel Xeon CPU for this application. In addition to this, we compared our fully-optimized code on the Intel Xeon Phi with a Matlab code running on single Intel Xeon CPU. Our method on the Intel Xeon Phi runs 16 times faster than the Matlab implementation. The result also suggests that the Intel Xeon Phi can offer an efficient but more general-purposed way to parallelize the deep learning algorithm compared to GPU. It also achieves faster speed with better parallelism than the Intel Xeon CPU.

关键词： Deep learning Unsupervised learning Deep architecture Sparse autoencoder Restricted Boltzmann Machine parallel algorithm Intel Xeon Phi Many-core

来源：评论

学校读者我要写书评

暂无评论

NUMERICAL STUDY OF FLOW AND HEAT TRANSFER IN ROTATING MICROCHANNELS

NUMERICAL STUDY OF FLOW AND HEAT TRANSFER IN ROTATING MICROC...

引用

作者： PRATANU ROY Texas A&M University

学位级别：博士

Investigation of fluid flow and heat transfer in rotating microchannels is impor- tant for centrifugal microfluidics, which has emerged as an advanced technique in biomedical applications and chemical separations. The centrifugal force and the Coriolis force, arising as a consequence of the microchannel rotation, change the flow pattern significantly from the symmetric profile of a non-rotating channel. A suc- cessful design of a centrifugal microfluidic device depends on effectively regulating these forces in rotating microchannels. Although a large number of experimental studies have been performed in order to demonstrate the applications of centrifugal microfluidics in various fields, a systematic study on the effect of rotation, chan- nel aspect ratio, and wall boundary conditions on the fluid flow and heat transfer phenomena in rotating microchannels has not yet been conducted. During the present study, pressure-based finite volume solvers in both staggered and collocated grids were developed to solve steady and unsteady, incompressible Navier-Stokes equations. The serial solver in collocated grid was parallelized using a Message Passing Interface (MPI) library. In order to accelerate the convergence of the collocated finite volume solver, a non-linear multi-grid method was developed. The parallel performances of the single and multi-grid solvers were tested on a two- dimensional lid driven cavity flow. High fidelity benchmark solution to a lid driven cavity flow problem in a 1024 × 1024 grid was presented for a range of Reynolds num- bers. parallel multigrid speedup as high as three orders of magnitude was achieved for low Reynolds number flows. In addition, the optimal multigrid efficiency was validated. The fluid flow in a rotating microchannel was modeled as a steady, laminar in- compressible flow with no slip and slip boundary conditions. For no slip boundary condition, critical values of parameters that determine the extent of the centrifugal force and th

关键词： Rotating Microchannels Finite Volume Method Multigrid algorithm parallel algorithm Collocated Grid Slip Boundary Condition Level Set Method Thesis

来源：评论

学校读者我要写书评

暂无评论

3D electromagnetic simulation of integrative antenna

3D electromagnetic simulation of integrative antenna

引用

16th International Symposium on Antenna Technology and Applied Electromagnetics (ANTEM)

作者： Peng, Kai Fang, Jinyong Jin, Song China Acad Space Technol Xian Space Payload Syst Innovat Ctr Xian 710100 Peoples R China Beijing Remote Sensing Inst Beijing Beijing Peoples R China

ISBN: (纸本)9781479922253

This paper realized an algorithm which had expedient modeling and high computer speed for large structures based on FDTD-PIC method. With this algorithm the microwave tubes with integrative structure or antenna could be simulated without worry of memory limited. It's important when the result of experiment was disagree with separate design. Finally with a simulation of a MILO and integrative antenna, the correctness of algorithm had been validated.

关键词： FDTD-PIC algorithm parallel algorithm integrative antenna

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：