A5034009645- Distributed and Parallel Computing Systems
- Parallel Computing and Optimization Techniques
- Scientific Computing and Data Management
- Advanced Data Storage Technologies
- Interconnection Networks and Systems
- Service-Oriented Architecture and Web Services
- Modular Robots and Swarm Intelligence
- Distributed systems and fault tolerance
- Embedded Systems Design Techniques
- Cloud Computing and Resource Management
- Matrix Theory and Algorithms
- Atomic and Molecular Physics
- Peer-to-Peer Network Technologies
- IoT and Edge/Fog Computing
- Electron and X-Ray Spectroscopy Techniques
- Mobile Agent-Based Network Management
- Advanced Database Systems and Queries
- Historical Studies in Latin America
- Business Process Modeling and Analysis
- Politics and Society in Latin America
- Advanced Chemical Physics Studies
- Mexican Socioeconomic and Environmental Dynamics
- Software System Performance and Reliability
- Semantic Web and Ontologies
- Latin American and Latino Studies
University of Utah
2020-2024
Indian Institute of Technology Ropar
2020-2024
Cardiff University
2014-2024
Swinburne University of Technology
2020-2024
Microsoft (Finland)
2016-2020
Indiana University Bloomington
2016-2020
Rutherford Appleton Laboratory
2020
Artistic Realization Technologies
2008-2015
University of Southern Mississippi
1991-2011
East Midlands Academic Health Science Network
2006
This report presents a methodology for measuring the performance of supercomputers. It includes 13 Fortran programs that total over 50,000 lines source code. They represent applications in several areas engi neering and scientific computing, many cases codes are currently being used by computational re search development groups. We also present PERFECT standard, set guidelines allow portability to types machines. Furthermore, we some measures method ology recording sharing results among...
This paper outlines the content and performance of ScaLAPACK, a collection mathematical software for linear algebra computations on distributed memory computers. The importance developing standards computational message passing interfaces is discussed. We present different components building blocks indicate difficulties inherent in producing correct codes networks heterogeneous processors. Finally, this briefly describes future directions ScaLAPACK library concludes by suggesting...
Gene expression is regulated by transcription factors that interact with cis -regulatory elements. Predicting these elements from sequence data has proven difficult. We describe here a successful computational search for direct in particular temporal-spatial pattern the Drosophila embryo, based on single well characterized enhancer model. The fly genome was searched to identify containing same combination of as those found Experimental evaluation results demonstrates our method can correctly...
The authors describe ScaLAPACK, a distributed memory version of the LAPACK software package for dense and banded matrix computations. Key design features are use versions Level 3 BLAS as building blocks, an object-oriented interface to library routines. square block scattered decomposition is described. implementation right-looking LU factorization algorithm on Intel Delta multicomputer discussed, performance results presented that demonstrate scalability algorithm.< <ETX...
Abstract Recent relativistic calculations on low energy elastic scattering of electrons from atoms are reviewed. Two aspects emphasized : mathematical transformations and numerical methods used in the solution Dirac equation summation partial wave series ; results calculations, particular behaviour cross sections spin polarizations for different atoms, energies incident electron, approximations. The effects screening nucleus by atomic electrons, electron exchange (including use approximate...
The Universal Description, Discovery and Integration (UDDI) is a specification for distributed Web-based information registries Web Services. UDDI allows HTTP-enabled business services to be published, subsequently searched, based on their interface. consists of three components: "white pages" hold basic contact identifiers company, "yellow enable companies listed industry categories (using standard taxonomies), "green record interface details how service invoked. however limited in scope -...
This article discusses the core factorization routines included in ScaLAPACK library. These allow and solution of a dense system linear equations via LU, QR, Cholesky. They are implemented using block cyclic data distribution, built de facto standard kernels for matrix vector operations (BLAS its parallel counterpart PBLAS) message passing communication (BLACS). In implementing routines, major objective was to parallelize corresponding sequential LAPACK BLAS, BLACS, PBLAS as building blocks,...
Abstract The paper describes Parallel Universal Matrix Multiplication Algorithms (PUMMA) on distributed memory concurrent computers. PUMMA package includes not only the non‐transposed matrix multiplication routine C = A ⋅ B, but also transposed routines T B , and for a block cyclic data distribution. perform efficiently wide range of processor configurations sizes. together provide same functionality as Level 3 BLAS xGEMM. Details parallel implementation are given, results presented runs...
First Page
article Free Access Share on A message passing standard for MPP and workstations Authors: Jack J. Dongarra Univ. of Tennessee, Knoxville KnoxvilleView Profile , Steve W. Otto Oregon Graduate Institute, Beaverton BeavertonView Marc Snir IBM T. Watson Research Center, Yorktown Heights, NY NYView David Walker Oak Ridge National Lab, Ridge, TN TNView Authors Info & Claims Communications the ACMVolume 39Issue 7July 1996 pp 84–90https://doi.org/10.1145/233977.234000Published:01 July...
This paper discusses the design of linear algebra libraries for high performance computers. Particular emphasis is placed on development scalable algorithms multiple instruction data (MIMD) distributed memory concurrent A brief description EISPACK, LINPACK, and LAPACK given, followed by an outline ScaLAPACK, which a version currently under development. The importance block-partitioned in reducing frequency movement between different levels hierarchical stressed. use such helps reduce message...
Calculations on electron impact excitation of the n=1 and n=2 states hydrogenic ions have been made for atomic numbers Z ranging from 2 to 100 in a Coulomb-Born approximation which Dirac wavefunctions are used instead more usual Schrodinger ones. All one-body relativistic effects included, two-body effects, mutual spin-orbit coupling but not magnetic or retardation effects. Relativistic found contribute 10% cross sections most transitions at Z=25 up factor 6 Z=100. The conservation total...
Differential cross sections and spin polarizations for 3·5-500 eV electrons scattered from mercury have been calculated by the numerical solution of relativistic Hartree-Fock equations, thus including effect exchange but ignoring distortion atom correlation effects. At upper end energy range inclusion produces small significant changes which bring results into very good agreement with experiment. lower energies some discrepancies appear, even at 3·5 qualitative is obtained.