Competence in High Performance Computing 2010

Name: Competence in High Performance Computing 2010 | Proceedings of an International Conference on Competence in High Performance Computing, June 2010, Schloss Schwetzingen, Germany
Brand: Springer
Price: 96.29 EUR
Availability: OnlineOnly

Proceedings of an International Conference on Competence in High Performance Computing, June 2010, Schloss Schwetzingen, Germany

Christian Bischof Heinz-Gerd Hegering Wolfgang E. Nagel Gabriel Wittum(Editor)

Springer (Publisher)

1st Edition

Published on 13. February 2012

X, 234 pages

E-Book

PDF with digital watermarking

System requirements

978-3-642-24025-6 (ISBN)

€96.29incl. 7% vat

System requirements

for PDF with digital watermarking

E-Book Single Licence

Available for download

Description

More details

Other editions

Content

Intro
Competence in High Performance Computing 2010
Foreword
Contents
HICFD: Highly Efficient Implementation of CFD Codesfor HPC Many-Core Architectures
1 Introduction
2 A Generic SIMD Preprocessor
2.1 The Vectorizing Source-to-Source Transformator Scout
2.2 Practical Results
3 Performance Analysis of SIMD Codes
4 Hybrid Shared-Memory/MPI Parallelization
4.1 Hybrid Multistage Shared-Memory/MPI Parallelization of the TAU Flow Solver
4.2 Hybrid Shared-Memory/MPI Parallelization of the TRACE Flow Solver
5 Scalable Algorithms for Many-Core Architectures
5.1 Partitioning for Block-Structured Grids
5.2 Preconditioning for the Linear TRACE Solvers
6 Conclusions
References
Effective Holistic Performance Measurement at Petascale Using IPM
1 Introduction
2 Performance Monitoring with IPM
2.1 IPM Design Principles
2.2 Performance Data Event Sources
2.3 Module Concept
2.4 Performance Data Output and Post-Processing
2.5 HTML Profiling Report
3 Efficiency of IPM
4 Related Work
5 Conclusions
References
The J-Cell Project: Developing a Scalable and Robust Runtime Environment for Heterogeneous Many-Core Processors
1 Introduction
2 System Overview
3 Object Location and Retrieval
4 Fully Decentralized Consensus
5 Redundancy
6 Related Work
7 Future Work
8 Conclusion
References
Pre-Stack PRO: A High-Performance Seismic Data Processing Solution
1 Introduction to Seismic Data
2 Introduction to Pre-Stack PRO
2.1 Gather Flattening
2.2 Filtering Algorithms
2.3 Amplitude Analysis
2.4 Working Modes
3 The Pre-Stack PRO Architecture
3.1 Client
3.2 Backend
3.3 Scalability and Performance
4 Fundamentals of Pre-Stack PRO
4.1 GPI
4.2 Render- and Visualization Engine
4.3 Fraunhofer Parallel File System
5 Summary
A New Scalable Monitoring Tool Using Performance Properties of HPC Systems
1 Introduction
2 Related Work
3 Agent Architecture and Communication
4 Performance Analysis
5 Information Reduction
6 Results
7 Conclusion and Future Work
References
Design and Implementation of Distributed Meshes in HiFlow3
1 Introduction
2 Motivation
3 Geometry and Topology
4 Abstract Interface Classes
5 Mesh Implementation
6 Cell Types
7 Refinement and Coarsening
8 Distributed Meshes
9 Future Work
10 Conclusion
References
Static and Dynamic Properties of Curved Vapour-Liquid Interfaces by Massively Parallel Molecular Dynamics Simulation
1 Introduction
2 Curved Fluid Interfaces in Equilibrium: Theory
3 Curved Fluid Interfaces in Equilibrium: MD Simulation
4 Homogeneous Vapour to Liquid Nucleation: Theory
5 Homogeneous Vapour to Liquid Nucleation: MD Simulation
6 Conclusion
References
Score-P: A Unified Performance Measurement Systemfor Petascale Applications
1 Introduction
1.1 Scalability Challenge
1.2 Interoperability Challenge
1.3 Redundancy Challenge
2 Project Overview and Goals
3 Background and Related Work
4 The Score-P Measurement System
5 The Open Trace Format 2
6 The CUBE Profiling Format Version 4
7 The Online Access Interface
8 The Opari2 Instrumenter
9 Interfacing with TAU
10 Future Work
References
Parallelization of an Edge- and Coherence-Enhancing Anisotropic Diffusion Filter with a Distributed Memory Approach Based on GPI
1 Introduction
2 Algorithmic Details
2.1 Anisotropic Diffusion Filter
2.2 Parallelization Strategy
2.3 Communication Patterns
3 Results
4 Summary and Conclusion
References
eeClust: Energy-Efficient Cluster Computing
1 Introduction
2 Software Components
2.1 VampirTrace and Vampir
2.2 Scalasca
2.3 GridMonitor
3 Project Plan
4 Results
4.1 SPECPower
4.2 SPEC MPI2007 Energy Consumption Characteristics
4.3 Recording Energy Consumption with VampirTrace
4.4 Influence of Arithmetic Instructions and Data Transfers on Energy Consumption
5 Ongoing Work
5.1 Energy Efficiency Benchmark
5.2 Scalasca Integration
5.3 Local Agent
5.4 GridMonitor Integration
6 Conclusion
References
An Efficient and Flexible Parallel FFT Implementation Based on FFTW
1 Introduction
2 Definitions and Assumptions
3 The Parallel Three-Dimensional FFT Algorithm
3.1 One-Dimensional Serial FFT Supported by FFTW
3.2 Two-Dimensional Parallel Transpositions Supported by FFTW
3.3 Parallel Three-Dimensional FFT Based on FFTW
4 Parallel FFT Software Library
5 Runtime Measurements
6 Concluding Remarks
References
ParaSCIP: A Parallel Extension of SCIP
1 Introduction
2 SCIP: Solving Constraint Integer Programs
2.1 Mixed Integer Programs
2.2 Branch-and-Bound
3 ParaSCIP
3.1 A Dynamic Load Balancing Mechanism
3.1.1 Initialization Phase
3.1.2 Transferring Branch-and-Bound Nodes
3.1.3 Load Balancing
3.2 Checkpointing and Restarting
4 Solving Open Instances from MIPLIB2003 on HLRN II
4.1 ds and stp3d Instances
4.2 Extended Presolving
4.3 HLRN II
4.4 Computational Results
5 Concluding Remarks
References
High-Performance Computing Techniques for Coupled Fluid, Structure and Acoustics Simulations
1 Introduction
2 Fundamental Equations and Model Approaches
2.1 Flow
2.2 Flow and Aeroacoustics
2.3 Structure
3 Numerical Realization
3.1 Basic Tools
3.2 Fluid-Structure Interaction
3.3 Flow-Acoustics-Coupling
4 Results
4.1 Fluid-Structure Interaction
4.2 Flow and Acoustics
5 Summary and Outlook
References
LIKWID: Lightweight Performance Tools
1 Introduction
2 Tools
2.1 likwid-perfctr
2.2 likwid-pin
3 Case Studies
3.1 Case Study 1: Influence of Thread Topology on STREAM Triad Performance
3.2 Case Study 2: Monitoring the Performance of a Lattice Boltzmann Fluid Solver
3.3 Case Study 3: Detecting ccNUMA Problems on Modern Compute Nodes
4 Conclusion and Future Plans
References
Scalability Issues of In-Situ Visualization in Parallel Simulation of Unsteady Flows
1 Introduction
2 Related Work
3 The DSVR Framework
4 Parallel Algorithms for Pathline Extraction
4.1 Seeding Strategies
4.2 Naive Parallelized Pathline Extraction
4.3 Advanced Parallelized Pathline Extraction
4.4 A Combined Algorithm
5 Evaluation
6 Conclusion and Future Work
References
Towards Intelligent Management of Very Large Computing Systems
1 Introduction
2 Related Work
3 TIMaCS: Solution
3.1 High Level Architecture
3.2 Monitoring
3.2.1 Data-Collector
3.2.2 Storage
3.2.3 Aggregator
3.2.4 Regression Tests
3.2.5 Compliance Tests
3.2.6 Filter & Event Generator
3.3 Management
3.3.1 Event Handler
3.3.2 Decision Maker
3.3.3 Knowledge Base
3.3.4 Controller, Controlled and Delegate
3.4 Virtualization in TIMaCS
3.5 Communication Infrastructure
4 Conclusion
References
Parallel Multigrid Methods and Coarse Grid LDLT Solver for Maxwell's Eigenvalue Problem
1 Introduction
2 The Discrete Maxwell Eigenvalue Problem
3 A Parallel Block LDLT Decomposition
4 Parallel Results
References
Scaling Algebraic Multigrid Solvers: On the Road to Exascale
1 Introduction
2 The Algebraic Multigrid Solver
2.1 Overview
2.2 Scaling Strategy for AMG
3 Smoothers
3.1 Hybrid Gauss-Seidel
3.2 Chebyshev
3.3 Numerical Results
4 On-node Threading and Memory Management
5 Scaling of AMG
6 Conclusions
References
Radiation Hydrodynamics Code LARED-H for Laser Fusion Simulation
1 Introduction
2 Multi-Block Structured Grids
3 Radiation Hydrodynamics
3.1 Hydrodynamics Equations
3.2 Radiation Energy Equations
4 Numerical Simulation of Ignition Target
5 Conclusions and Perspectives
References

System requirements

Save as PDF Copy link into clipboard

Schweitzer Fachinformationen

Competence in High Performance Computing 2010

Description

More details

Other editions

Additional editions

Content

System requirements