Conference on Hypercube Concurrent Computers and Applications最新文献_第3页

Distributed orthogonal factorization 分布正交分解

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63122

A. Pothen, P. Raghavan

引用次数: 39

Large-scale concurrent computing in artificial intelligence research 大规模并发计算在人工智能研究中的应用

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63089

L. Gasser

{"title":"Large-scale concurrent computing in artificial intelligence research","authors":"L. Gasser","doi":"10.1145/63047.63089","DOIUrl":"https://doi.org/10.1145/63047.63089","url":null,"abstract":"Research in AI is slowly maturing, and body of accepted techniques for reasoning and for representing knowledge in simple, circumscribed domains now exists. But with the maturity of AI has come a growing awareness of the severe limitations of current techniques for constructing more complex problem solving or interpretation systems. We currently have inadequate means to gather, represent, store, organize, access, and manipulate the huge collections of knowledge required for complex problem solving. Existing systems can't reconfigure themselves in changing situations, nor can they incrementally adjust to new knowledge or new techniques. Large scale problem solvers (e.g. factory automation systems) cannot in principle completely model the world in which they exist, and must face problems of inconsistency, asynchrony, control and geographic distribution, etc. — they will have to work in “open systems.”\u0000Many solutions under consideration rely on concurrent computation, using either very fine grained “connectionist,” “neural computing” or “data parallel” approaches, or using larger grain collections of “objects,” “agents,” or “problem solving nodes” — techniques collectively termed “Distributed AI.” In this paper we characterize the needs for concurrency and parallelism in AI, with special attention to building medium to large grain adaptive problem solvers in open systems. In these systems the overriding concern is organizing the problem solving system's behavior — the “coordination problem.” Conventional distributed computing and parallel algorithms approaches allow a programmer to solve the coordination problem, and provide language constructs and concurrency control mechanisms with which a program can enact his solution. In Distributed AI, we attempt to improve adaptability by designing problem solvers which can both solve the coordination problem and enact the solution themselves.","PeriodicalId":299435,"journal":{"name":"Conference on Hypercube Concurrent Computers and Applications","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1989-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116768341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 12

A hybrid hypercube algorithm for the symmetric tridiagonal eigenvalue problem 对称三对角特征值问题的混合超立方体算法

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63113

J. A. Jackson, L. Liebrock, L. Ziegler

引用次数: 1

Vectorized dissection on the hypercube 超立方体上的矢量解剖

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63135

T-H. Olesen, J. Petersen

引用次数: 4

Pattern recognition by neural network model on hypercubes 超立方体上的神经网络模式识别

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63055

W. Furmanski

引用次数: 8

The prime factor non-binary discrete Fourier transform and use of Crystal_Router as a general purpose communication routine 素因子非二进制离散傅里叶变换和使用Crystal_Router作为通用通信例程

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63087

G. Aloisio, Nicola Veneziani, Jai Sam Kim, G. Fox

引用次数: 5

Block-matrix operations using orthogonal trees 使用正交树的块矩阵操作

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63115

A. Elster, A. Reeves

{"title":"Block-matrix operations using orthogonal trees","authors":"A. Elster, A. Reeves","doi":"10.1145/63047.63115","DOIUrl":"https://doi.org/10.1145/63047.63115","url":null,"abstract":"Hypercube algorithms are presented for distributed block-matrix operations. These algorithms are based entirely on an interconnection scheme which involves two orthogonal sets of binary trees. This switching topology makes use of all hypercube interconnection links in a synchronized manner.\u0000An efficient novel matrix-vector multiplication algorithm based on this technique is described. Also, matrix transpose operations moving just pointers rather than actual data, have been implemented for some applications by taking advantage of the above tree structures. For the cases where actual physical vector and matrix transposes are needed, possible techniques, including extensions of the above scheme, are discussed.\u0000The algorithms support submatrix partitionings of the data, instead of being limited to row and/or column partitionings. This allows efficient use of nodal vector processors as well as shorter interprocessor communication packets. It also produces a favorable data distribution for applications which involve near neighbor operations such as image processing. The algorithms are based on an interprocessor communication paradigm which involves variable length, tagged block data transfers. They have been implemented on an Intel iPSC hypercube system with the support of the Hypercube Library developed at the Christian Michelsen Institute.","PeriodicalId":299435,"journal":{"name":"Conference on Hypercube Concurrent Computers and Applications","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1989-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124579529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 12

The Flux-corrected transport algorithm on the NCUBE hypercube NCUBE超立方体上的通量校正传输算法

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63065

D. Walker, G. Fox, G. Montry

引用次数: 1

A universal concurrent algorithm for plasma particle-in-cell simulation codes 等离子体细胞内粒子模拟代码的通用并行算法

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63063

P. Liewer, V. Decyk, J. Dawson, G. Fox

{"title":"A universal concurrent algorithm for plasma particle-in-cell simulation codes","authors":"P. Liewer, V. Decyk, J. Dawson, G. Fox","doi":"10.1145/63047.63063","DOIUrl":"https://doi.org/10.1145/63047.63063","url":null,"abstract":"We have developed a new algorithm for implementation of plasma particle-in-cell (PIC) simulation codes on concurrent processors. This algorithm, termed the universal concurrent PIC algorithm (UC-PIC), has been utilized in a one-dimensional electrostatic PIC code on the JPL Mark III Hypercube parallel computer. To decompose the problem using the UC-PIC algorithm, the physical domain of the simulation is divided into sub-domains, equal in number to the number of processors, such that all sub-domains have roughly equal numbers of particles. For problems with non-uniform particle densities, these sub-domains will be of unequal physical size. Each processor is assigned, a sub-domain, with nearest neighbor sub-domains assigned to nearest neighbor processors. Using this algorithm in the Mark III PIC code, the increase in speed in going from 1 to 32 processors for the dominant portion of code (push time, defined below) was 29, yielding a parallel efficiency of 90%. Although implemented on a hypercube concurrent computer, this algorithm should be also be efficient for PIC codes on other parallel architectures and on sequential computers where part of the data resides in external memory.","PeriodicalId":299435,"journal":{"name":"Conference on Hypercube Concurrent Computers and Applications","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1989-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116780463","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 13

A concurrent neural network algorithm for the traveling salesman problem 旅行商问题的并行神经网络算法

Conference on Hypercube Concurrent Computers and Applications Pub Date : 1989-01-03 DOI: 10.1145/63047.63105

N. Toomarian

引用次数: 7