检索结果-内蒙古大学图书馆

General implementation of multilevel parallelization in a gradient-based design optimization algorithm

AIAA JOURNAL 2006年第9期44卷 1993-2001页

作者： Rajan, S. D. Belegundu, A. D. Damle, A. S. Lau, D. St Ville, J. Arizona State Univ Dept Civil & Environm Engn Tempe AZ 85287 USA Penn State Univ Mech & Nucl Engn Dept University Pk PA 16802 USA Hawthorne & York Int FEA Grp Phoenix AZ 85008 USA

As product designs have become more sophisticated, both the simulation models (e.g., finite element models) and the design optimization models have grown bigger. To keep pace with this increase in problem size, we present and implement an optimization strategy that can run on a computing cluster with demonstrable efficiency. First, parallelism is implemented in the context of gradient calculations using divided differences. Then, parallelism is achieved in the context of both direction-finding and line-search steps. parallel direction finding improves the convergence rate as opposed to just cutting down the amount of arithmetic. A new algorithm based on method of feasible directions is discussed that obtains better optima and is also computationally faster. Implementation details regarding distribution of computing tasks to improve scalability and load balancing are presented. Numerical examples show the efficiency of the developed methodology on a relatively small computing cluster. Gains of about 7:1 have been obtainable using 16 processors on some test problems. Importantly, the framework presented can be developed by researchers using other gradient-based optimization codes on different computing platforms.

关键词： Sequential Quadratic Programming Finite Element Analysis parallel computing architectures Numerical Optimization Planar Truss parallel Algorithm parallel computing Supercomputers Software Development parallel Computers

来源：评论

学校读者我要写书评

暂无评论

Virtual signal processors

引用

MICROPROCESSORS AND MICROSYSTEMS 1998年第3-4期22卷 135-148页

作者： Dick, C Harris, F La Trobe Univ Sch Elect Engn Bundoora Vic 3083 Australia San Diego State Univ Dept Elect & Comp Engn San Diego CA 92182 USA

The introduction of SRAM-based field programmable gate arrays (FPGAs) has opened up a new dimension to parallel computing architectures. This paper describes an alternative approach to parallel computing - reconfigurable or virtual parallel processing (VPP). Rather than mapping an application onto a given parallel machine, the WP approach synthesizes the appropriate type and number of processing elements, as well as the interconnection topology, that is optimal for the application. For each application, configuration data is downloaded to the machine that personalizes the hardware for the task at hand. The paper provides a brief description of the authors reconfigurable computer, Archimedes. The benefits of the VPP approach are highlighted by two example applications - the 2-D FFT and a narrow-band digital filter. A novel parallel implementation of a polynomial transform based 2-D transform is described and compared with results for distributed memory parallel machines that have been reported in the literature. The comparison highlights the computational advantage provided by reconfigurable computing. The digital filter implementation employs sigma-delta modulation encoding to reduce the arithmetic workload. The application of this technology to a communications receiver is explained. (C) 1998 Elsevier Science B.V.

关键词： parallel computing architectures FPGA reconfigurable computing FFT polynomial transform

来源：评论

学校读者我要写书评

暂无评论

A parallel QUASI-LINEARIZATION ALGORITHM FOR AIR VEHICLE TRAJECTORY OPTIMIZATION

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 1986年第1期9卷 119-121页

作者： MENON, PKA LEHMAN, LL Integrated Syst. Inc. Palo Alto CA USA

The quasi-linearization algorithm using the integrating matrix approach described in this paper reduces the nonlinear Two-point boundary-value problem to a series of linear algebraic systems and is highly suited for implementation on parallel computing machines. The efficacy of this algorithm was illustrated for a boost-glide missile trajectory optimization on a serial computing machine. This approach can also be used for solving multipoint boundary-value problems and systems with state-costate discontinuities. Mechanization of this algorithm on state-of-the-art parallel processors will be of future interest.

关键词： Transversality Condition Flight Path Angle Nonlinear Programming Boost Glide Aerodynamic Drag Boundary Value Problems Dynamic Pressure parallel computing architectures Airspeed Richardson Extrapolation

来源：评论

学校读者我要写书评

暂无评论

NONRECURSIVE ORDER-N FORMULATION OF MULTIBODY DYNAMICS

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 1993年第5期16卷 838-844页

作者： KURDILA, AJ MENON, RG SUNKEL, JW NASA LYNDON B JOHNSON SPACE CTRDIV NAVIGAT CONTROL & AERONAUTHOUSTONTX 77058

Although excellent progress has been made in deriving algorithms that are efficient for certain combinations of system topologies and concurrent multiprocessing hardware, several issues must be resolved to incorporate transient simulation in the control design process for large space structures. Specifically, strategies must be developed that are applicable to systems with numerous degreees of freedom. Algorithms are required that induce parallelism on a fine scale, suitable for the emerging class of highly parallel processors. This paper addresses these problems by employing the range space formulation of multibody dynamics and solving for multipliers using the preconditioned conjugate gradient method. By employing regular ordering of the system connectivity graph, an extremely efficient preconditioner can be derived from the range space metric. The method can achieve performance rates that depend linearly on the number of substructures. Furthermore, the approach is promising as a potential parallel processing algorithm in that it exhibits fine parallel granularity, is nonassembling, and is easily load balanced among processors without relying an system topology to induce parallelism.

关键词： Multibody Dynamics Conjugate Gradient Method Large Space Structures Space Stations Conjugate Gradient Algorithm Trusses Finite Element Modeling Multibody Simulation Transient Response parallel computing architectures

来源：评论

学校读者我要写书评

暂无评论

A Simulation Architecture for Self-Assembling and Reconfigurable Space Structures

A Simulation Architecture for Self-Assembling and Reconfigur...

引用

AIAA Science and Technology Forum and Exposition (AIAA SciTech Forum)

作者： Battista, Thomas A. Helvajian, Henry Aerosp Corp Control Anal Dept El Segundo CA 90245 USA Aerosp Corp Phys Sci Labs El Segundo CA 90245 USA

ISBN: (数字)9781624106095

ISBN: (纸本)9781624106095

We present a simulation architecture for a dynamic system composed of modular and interconnected small satellites. Standard modeling approaches rooted in analytical mechanics do not readily admit changes to the configuration space of a dynamic system. However, the proposed architecture is designed to handle discrete connection and detachment events using a dynamic constraint management algorithm. The capability of the simulation architecture is demonstrated using HIVE, a modular space architecture.

关键词： Space Structures Space Architecture Small Satellites Kinetic Energy James Webb Space Telescope Reaction Wheels parallel computing architectures Holonomic Constraints Inverse Kinematics Control Algorithm

来源：评论

学校读者我要写书评

暂无评论

Interoperable Application Programming Interfaces for Computer Aided Engineering Applications

Interoperable Application Programming Interfaces for Compute...

引用

AIAA Science and Technology Forum and Exposition (AIAA SciTech Forum)

作者： Jones, William T. Wood, Stephen L. Jacobson, Kevin E. Anderson, W. Kyle NASA Langley Res Ctr Computat AeroSci Branch MS 128 Hampton MS 23681 USA NASA Langley Res Ctr Aeroelastic Branch MS 340 Hampton MS 23681 USA

ISBN: (数字)9781624106095

ISBN: (纸本)9781624106095

A suite of software interfaces has been developed to facilitate the coupling of disciplines of interest to Computer Aided Engineering. The definition of the interfaces focused on abstraction and interoperability between components of coupled disciplines. The interfaces are language independent and allow for instantiable state to be maintained by the respective implementation. However, to support the largest spectrum of programming models, C calling convention has been adopted. The use of software interfaces places the dependency on the interface and not on the specifics of its implementation. Therefore, different implementations of a given interface can be seamlessly adopted as needed. Implementation of the interfaces remains the responsibility of component developers who have total control over the details of fulfillment. However, strict adherence to the interface definition guarantees interoperability between components. The interface definition is successfully demonstrated through application to Computational Fluid Dynamics.

关键词： Computer Aided Engineering Application Programming Interface Computational Fluid Dynamics FUN3D parallel computing architectures Data Structures Finite Volume Method Fluid Structure Interaction Graphical User Interface Finite Element Solution

来源：评论

学校读者我要写书评

暂无评论

MATE Network Tearing Techniques for Multiprocessor Solution of Large Power System Networks

MATE Network Tearing Techniques for Multiprocessor Solution ...

引用

IEEE-Power-and-Energy-Society General Meeting

作者： Tomim, Marcelo A. Marti, Jose R. De Rybel, Tom Wang, Lei Yao, Michael Univ Fed Juiz de Fora Juiz De Fora Brazil Univ British Columbia Vancouver BC Canada Powertech Labs Inc Surrey BC Canada British Columbia Transmiss Cooporat Victoria BC Canada

ISBN: (纸本)9781424483570

The Multi-Area Thevenin Equivalent algorithm (MATE) partitions a network into a number of subnetworks connected by links. The solution proceeds in two stages: first the subnetworks are solved independently, with the links open, and then the system formed by the links is solved with the subnetworks represented by their Thevenin equivalents as seen from the link nodes. In this paper, an overview of the MATE technique for solving large power systems on parallel computing architectures will be discussed, along with a performance analysis and applications, including the solution of the WECC system (about 15,000 nodes and 17,000 branches), which was achieved with a 7-time speedup with respect to a sequential sparsity-oriented algorithm. Comparisons with another parallel sparse solver are also presented.

关键词： MATE network tearing techniques WECC system large power system networks link nodes multiarea TheÌvenin equivalent algorithm multiprocessing systems multiprocessor solution parallel architectures parallel computing architectures parallel sparse solver power engineering computing power systems sequential sparsity-oriented algorithm

来源：评论

学校读者我要写书评

暂无评论

Mission Planning for Minimal Orbital Conjunctions

Mission Planning for Minimal Orbital Conjunctions

引用

AIAA SciTech Forum

作者： Rockwood, Troy D. Jah, Moriba K. Cheng, Yang Mitre Corp Colorado Springs CO 80910 USA Univ Texas Austin Austin TX 78705 USA XAnalytix Syst LLC Dept Aerosp Engn & Engn Mech Clarence Ctr NY 14032 USA

ISBN: (数字)9781624107115

ISBN: (纸本)9781624107115

Selection of destination orbits during mission planning involves a number of factors. Mission goals, the economic factors of size, weight, and power, and the operational constraints imposed by on-orbit conjunctions each affect orbit selection decisions. The number of conjunctions that a satellite will experience dictates mission lifetime and operational efficiency. Minimizing the effect of conjunctions involves significant analytic effort. An optimization tool is presented for mission planning. It takes mission-derived parameters such as acceptable altitudes and inclinations, and finds orbits with the minimum number of conjunctions with resident space objects utilizing the space object catalog. Global optimization techniques and implementation details are presented.

关键词： Resident Space Object Satellites Optimization Algorithm parallel computing architectures Orbital Capacity Sun Synchronous Orbit Space Environment Right Ascension Orbital Debris Propellant

来源：评论

学校读者我要写书评

暂无评论

Implementation of a Highly-parallel Finite-volume Test-Bench Code in OpenCL

Implementation of a Highly-Parallel Finite-volume Test-Bench...

引用

AIAA Aviation Forum

作者： Kedward, Laurence J. Allent, C. B. Computat Aerodynam Bristol BS8 1TR Avon England

ISBN: (数字)9781624105982

ISBN: (纸本)9781624105982

The trend for increased processor core numbers in both consumer and server CPUs, as well as the increasing use of many-core GPU accelerators for computation, has lead to a new paradigm of massive parallelism. GPUs in particular offer significant advantages in terms of memory bandwidth and floating-point performance as well as energy efficiency and price-to-performance. To leverage the new class of highly-parallel CPUs and GPUs, and close the gap between HPC capability and realised performance, numerical methods developed for the traditional multicore model must be reformulated for and combined with fine-grain thread-level parallelism. In this paper, background is given for computation on many-core architectures, with focus given to finite-volume methods for CFD. In particular, the memory-bound nature of finite-volume kernels requires significant attention to be given to code structure and optimisation. A finite-volume test-bench code is implemented in Fortran and OpenCL and the effect of arithmetic intensity on overall performance is demonstrated. Notably, the optimised code structure is very different from that of the original, with significant increases in performance being achieved through kernel merging. Preliminary results show an overall speed-up of between 18x and 24x on two high performance consumer GPUs compared to 16 threads of a ninth generation Intel CPU. Up to 90% of the peak memory bandwidth is achieved for single-grind computation, while many options still remain for further code optimisation.

关键词： Graphics Processing Unit Finite Volume Method Computational Fluid Dynamics FORTRAN Energy Efficiency Primitive Variable Kernel Functions Multigrid Methods parallel computing architectures Unstructured Grid

来源：评论

学校读者我要写书评

暂无评论

Fully-Kinetic PIFE-PIC Simulations of Plasma Charging at Lunar Craters

Fully-Kinetic PIFE-PIC Simulations of Plasma Charging at Lun...

引用

AIAA Science and Technology Forum and Exposition (AIAA SciTech)

作者： Lund, David Zhao, Jianxun Lamb, Alexander Han, Daoru Missouri Univ Sci & Technol Dept Mech & Aerosp Engn 400 W 13th St Rolla MO 65409 USA

ISBN: (数字)9781624105951

ISBN: (纸本)9781624105951

This paper presents a fully-kinetic numerical investigation of plasma charging at lunar craters using the most recently developed parallel Immersed-Finite-Element Particle-in-Cell (PIFE-PIC) code. This model explicitly includes the lunar regolith layer and the bedrock in the simulation domain, taking into account the regolith layer thickness and permittivity, and is capable of resolving a non-trivial surface terrain or spacecraft configuration. 3-D domain decomposition is used for both particle-push and field-solve steps of the PIC loop to distribute the computation among multiple processors. Simulations are presented to study surface charging and lunar habitat charging near the lunar craters.

关键词： Permittivity Spacecraft Configurations Solar Wind Electron Density parallel computing architectures Dirichlet Boundary Condition Plasma Surface Interaction Electrostatic Fields Lunar Exploration Jet Propulsion Laboratory Solar Radiation

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：