High Performance Cluster Computing

High Performance Cluster Computing : Programming and Applications, Volume 2

4 (11 ratings by Goodreads)
By (author) 

List price: US$55.51

Currently unavailable

Add to wishlist

AbeBooks may have this title (opens in new window).

Try AbeBooks

Description

A comprehensive guide to todays most advanced R&D in highly parallel programming and applications. Volume 1 of this two-volume set collected todays best work on the systems aspects of high performance cluster computing. Now, in High Performance Cluster Computing: Programming and Application Issues, Volume 2, Rajkumar Buyya brings together the worlds leading work on programming and applications for todays state-of-the-art commodity supercomputers. The book is organized into three areas: programming environments and development tools; Java(tm) as a language of choice for development in highly parallel systems; and state-of-the-art high performance algorithms and applications. All three areas have seen major advances in recent years-and in all three areas, this book offers unprecented breadth and depth. Coverage includes: * New parallel programming techniques and tools, including MP and PVM, active objects, scoped behavior, and LiPS. State-of-the-art debuggng techniques: Code liberation, global renaming, name reclamation, and debugging interfaces. The WebOS: Designing operating system services for wide-area applications. Leveraging Java(tm) to the fullest: Distributed objects, the H
show more

Product details

  • Hardback | 664 pages
  • 175.26 x 236.22 x 33.02mm | 1,156.65g
  • Prentice Hall
  • Upper Saddle River, United States
  • English
  • 0130137855
  • 9780130137852

Table of contents

Preface.
I. PROGRAMMING ENVIRONMENTS AND DEVELOPMENT TOOLS.

1. Parallel Programming Models and Paradigms.


Introduction. A Cluster Computer and its Architecture. Parallel Applications and Their Development.



Strategies for Developing Parallel Applications.



Code Granularity and Levels of Parallelism. Parallel Programming Models and Tools.



Parallelizing Compilers. Parallel Languages. High Performance Fortran. Message Passing. Virtual Shared Memory. Parallel Object-Oriented Programming. Programming Skeletons.



Methodical Design of Parallel Algorithms.



Partitioning. Communication. Agglomeration. Mapping.



Parallel Programming Paradigms.



Choice of Paradigms. Task-Farming (or Master/Slave). Single-Program Multiple-Data (SPMD). Data Pipelining. Divide and Conquer. Speculative Parallelism. Hybrid Models.



Programming Skeletons and Templates.



Programmability. Reusability. Portability. Efficiency.



Conclusions. Bibliography.

2. Parallel Programming Languages and Environments.


Introduction. Early Mechanisms.



Semaphores. Conditional Critical Regions. Monitors. Sockets. Remote Procedure Calls. Rendezvous.



Shared Memory Environments.



Pure Shared Memory Environments. Virtual Shared Memory Environments.



Distributed Memory Environments.



Ada. Message Passing Interface. Parallel Virtual Machine. Distributed Computing Environment. Distributed Java.



Parallel Declarative Environments.



Parallel Logic Languages. Parallel Functional Languages.



Summary. Bibliography.

3. MPI and PVM Programming.


Introduction. Comparison of MPI and PVM. The All Pairs Shortest Path Problem.



Description of the Problem. Dijkstra's Algorithm. Floyd's Algorithm. Parallel Algorithms.



The MPI Programming Environment.



Communication Models in MPI. Creating an MPI Program. Running an MPI Program. PDIJK SVIPI. PFLOYD_MPI. Tools for MPI Programs.



The PVM Programming Environment.



Communication Models in PVM. Creating a PVM Program. Running a PVM Program. PDIJK AVM. P FLOYD_PVM. Tools for PVM Programs.



Porting Hints.



Initial Environment Setup. Parallel Tasks Setup. Group Management. Intertask Communication. Utility Functions.



Summary. Bibliography.

4. Linking Message-Passing Environments.


Interoperability Between Message-Passing Interfaces.



Message Passing Between Programming Environments. Access to Resource Management Systems.



An Overview of the PLUS Library.



Process Creation and Management.



System Architecture.



Library. Daemons.



Adding New Message-Passing Environments. Performance Results. Related Work. Summary. Bibliography.

5. Active Objects.


Objects in Cluster-Based Parallel Systems. Active Versus Passive Objects. Objects and Atomicity. BaLinda K Objects. Speculative Processing. Summary. Bibliography.

6. Using Scoped Behavior to Optimize Data Sharing Idioms.


Introduction. Motivation: Data Sharing Idioms. Aurora: A Distributed Shared Data System.



Process Models. Shared-Data Objects. Scoped Behavior. Illustrative Example: Matrix Multiplication. Discussion: Programming in Aurora.



Implementation Overview.



Handle-Body Composite Objects. Scoped Behavior Objects.



Experience with Parallel Programs.



Matrix Multiplication. 2-D Diffusion. Parallel Sorting by Regular Sampling.



Discussion and Related Work. Concluding Remarks. Bibliography.

7. Component-Based Development Approach.


Introduction. Component-Based Application Development.



Overall Framework. Programmer Interface. Building Reusable Design Components. Assembling Reusable Design Components. Compiling and Running a Distributed Application. Alternative Design Styles. Consistency Checks.



Advanced Features.



Architecture Models as Reusable Components. Generation of Communication and Synchronization Code. Manipulating Messages. Repetitive Substructures in Architecture Models.



Reusing Simulation Software in a Distributed Setting. Comparison Between Approaches. Concluding Remarks. Bibliography.

8. Hypercomputing with LiPS.


Generative Communication.



Tuple Space. Benefits of Generative Communication. Example Applications.



Using LiPS.



Writing and Generating a LIPS Application. Starting the System and its Application. Monitoring the System and its Applications. Some Remarks on Fault-Tolerant Applications.



The LIPS Runtime Systems.



The Fault-Tolerant Tuple Space Machine. Performance. The LIPS Daemon (lipsd()). The LIPS Daemon Controller (lipsdc()).



The LIPS Development System.



Programming in CWEB. The LIPS Test Environment.



Bibliography.

9. An Efficient Tuple Space Programming Environment.


Introduction. Tuple Space Programming.



Fundamentals. Example Linda Program. Associative Memory Analysis.



Compilation Environment.



Basic Translation. Optimizing Compilers.



Run-time Environment.



Processor Location of Data. Data Structures for Efficient Data Access. Data Transfer Protocol. Process Creation. Cluster Execution Environment. Run-time Optimizations.



Extensions. Conclusions. Bibliography.

10. Debugging Parallelized Code.


Introduction. Automatic Parallelization. The Debugging Problem. Debugging with Code Liberation.



Overview of the Technique. The Single Assignment Form via Global Renaming. Renaming Simple Variables in Structured Code. Array Renaming. Renaming Unstructured Code. Name Reclamation. Reclaiming the Names. The Debugging Interface.



Experimental Results. Conclusions. Bibliography.

11. WebOS: Operating System Services for Wide-Area Applications.


Introduction. WebOS Overview. Naming. Persistent Shared State. Security and Authentication.



Validating and Revoking Statements. Processes and Roles. Authorization.



Process Control. Rent-A-Server.



Current Approaches. System Design. Performance.



Related Work. Conclusions. Bibliography.

II. JAVA FOR HIGH PERFORMANCE COMPUTING.

12. Distributed-Object Computing.


Introduction. CORBA.



Basic Model - C ORBA 2.0 Architecture. OMG IDL(Interface Definition Language) and its Mapping. C ORBA Object Services. A Matrix Multiplication Example.



Java RMI.



Basic Model. RMI Features. Matrix Multiplication.



DCOM.



Basic Model. Identification of DCOM Interfaces and Classes. IDL and its Mapping. Creation of Objects. Using COM Interfaces and Object Lifetime. DCOM Programming in Java. Matrix Multiplication Example.



Voyager.



Basic Model. Voyager Features.



A Simple Performance Measurement.



Matrix Multiplication. Which Paradigm is Superior?



Bibliography.

13. Java and Different Flavors of Parallel Programming Models.


Introduction. Java Threads-Built-in Support for Parallelism and Concurrency.



Are Java Threads the Correct Model? Java Support for Parallelism.



Parallel Programming Models.



Functional Models. Object-Oriented Models. Data Parallel Models. Message Passing Models. Shared Memory Models.



Summary. Bibliography.

14. The HPspmd Model and its Java Binding.


Introduction. Java Language Binding.



Basic Concepts. Global Variables. Program Execution Control. Communication Library Functions.



Java Packages for HPspmd Programming. Programming Examples. Issues in the Language Design.

15. Web-Based Parallel Computing with Java.


Introduction. Web-Based Parallel Computing. Comparing Cluster with Web-Based Parallel Computing. Examples of Internet-Based Parallel Computing. Can Java be Used for Web-Based Parallel Computing? Problems to be Solved in Web-Based Parallel Computing. A Case Study: The JET Platform. Some Performance Results. Conclusions. Bibliography.

III. ALGORITHMS AND APPLICATIONS.

16. Object-Oriented Implementation of Parallel Genetic Algorithms.


Introduction. Short Overview of GA Systems. Object-Oriented Approach to PGAs. Classes Representing Individuals. Local Genetic Operations. Island Model. Global Population Model. Load Balancing. File and I/O Operations. PGA Application Framework. Sample Results. Concluding Remarks. Bibliography.

17. Application-Specific Load Balancing on Heterogeneous Systems.


Introduction. System Overview. Implementation of a Complex FDTD Equation.



Implementation Details. Experimental Result.



Load Balancing.



Methodology. Processor Weights. Application Data Movement (ADM).



Analysis.



Perfect Load Balancing System. Experimental Results. Future Improvements.



Conclusions. Bibliography.

18. Time Management in Parallel Simulation.


Introduction.



Different Types of Simulations.



Major Issues of Parallel Simulation. Principles of Parallel Simulation.



Distributed Programming: A Simulation Example. The Apparently Sequential Nature of Simulation. A Natural Correspondence Between Problem and Solution. Logical Process Simulation. Conservative Vs. Optimistic Simulation.



Conservative Synchronization Protocols.



Variants of the Chandy-Misra Null Message Protocol. Deadlock Detection and Recovery. Global Vs. Local Lookahead. Conservative Time Windows.



Conclusion. Bibliography.

19. Hardware System Simulation.


Introduction. NEPSi.



ASIMUT. NEPSi as a Client Server Architecture. Overview of the NEPSi Network. Communication Model. VHDL Models Used. Tests for TNP. Initial Tests. Varying Numbers of PEs and Simulators. Matrix Multiplication on TNP. Variable Machines, Processes and PEs. Rollback. Multi Campus Simulation. Effect of Migration / Load Balancing. Network Traffic Effects.



Discussion. Bibliography.

20. Real-Time Resource Management Middleware: Open Systems and Applications.


Introduction. Architecture of Dynamic (aoS Management Middleware. Adaptive Resource Allocation.



Automatic Survivability. Automatic Scalability. Assessment of Real-Time QoS. Resource Allocation and Management.



Experiences with the Adaptive Resource Management Services.



The Navy Testbed. A Real-Time Control System Benchmark. Experiments.



Conclusions. Bibliography.

21. Data Placement in Shared-Nothing Database Systems.


Introduction. Data Placement. Declustering.



Evaluation Time Reduction for Select and Project. Relation CoLocation for Join.



Placement. Re-Distribution.



Update Re-Distribution. Query Frequency Change Re-Distribution.



Dynamic Re-Organization. Summary. Bibliography.

22. Parallel Inference with Very Large Knowledge Bases.


Introduction.



Parallel Reflexive Reasoning on Cluster Computers. Contributions.



SHRUTI: A Structured Connectionist Reasoning System.



Terminology. Knowledge Encoding and Inference in Backward Reasoning. Constraints on Rules and Inferences.



Mapping SHRUTI onto Parallel Machines. SHRUTI on the CM-5-Design and Implementation.



The Connection Machine CM-5. The SHRUTI-CM5 Knowledge Base. Design Considerations. Encoding the Knowledge Base. Spreading Activation and Inference.



SHRUTI-CM5-A Mathematical Analysis.



Motivation. A Summary of the Analysis. Implications of the Analysis.



SHRUTI-CM5-Experiments with Large Knowledge Bases.



Experiments with Random Knowledge Bases. Experiments with WordNet. Empirical Validation of the Analysis.



Conclusion. Bibliography.

23. MaRT: Lazy Evaluation for Parallel Ray Tracing.


Introduction.



Principles of Ray Tracing. Accelerated Ray Tracing Techniques. The Time-Memory Balance.



On Ray Tracing Parallelization Techniques.



The Image Parallel Approach. The Object Parallel Approach. Parallel Computation Analysis. Synthesis.



MaRT: a Lazy Ray Tracer.



On Lazy Evaluation. The Lazy Octree. Lazy Construction of Polygon Data.



Parallel MaRT.



Implementation of Parallel MaRT. Results.



Concluding Remarks. Bibliography.

24. Fast Content-Based Image Retrieval.


Introduction. Image Feature Extraction.



AnOverview. Wavelet Based Multiple Image Feature Extraction.



Dynamic Image Indexing. Image Similarity Measurement. Image Searching. Parallel Implementation.



Divide-and-Conquer Parallel Algorithms.



Experimental Results.



Parallel Wavelet Transform Using PVM.



Parallel Image Feature Extraction - PVM Vs. DSM.



Hierarchical Image Matching in a PVM Environment.



Conclusion. Bibliography.

25. Climate Ocean Modeling.


Introduction. Model Description. Parallel Partition on Irregular Geometries.



A Simple Partition. An Efficient Partition.



Ocean Modeling on Various Systems.



Overview of Parallel Systems. Cray T3D. Cray T3E. HP SPP2000. Beowulf System-PC Clusters. Code Performance and Intercomparisons.



Scientific Results. Summary. Bibliography.

26. Computational Electromagnetics.


Introduction. Physical Optics Method. Finite-Difference Time-Domain Method. Finite-Element Integral-Equation Coupled Method. Conclusions. Bibliography.

27. CFD Simulation: A Case Study in Software Engineering.


Introduction. TfC - a State-of-the-Art Industrial CFD Package. Requirements for Parallel CFD Simulation. Design and Implementation of ParTfC. Object Oriented Design of Scientific Software. Productive Use of Parallel Scientific Computing Software.



Resource Management in Workstation Clusters. The Pre- and Post-Processing Bottlenecks. Network Based Collaborative Design Environments.



Bibliography.

28. Quantum Reactive Scattering Calculations.


Introduction. The Many Body Problem: Description, Decomposition, and Solutions.



The Decomposition of the Problem and the Mathematical Formalism. The Integration of Scattering Equations.



Parallelization Strategies.



Task Farm Model. Pipeline Model.



Parallel Implementations on CRAY T3E.



Task Farm Model. Pipeline Model. Performance Evaluation.



Parallel Implemention on SGI Origin 2000.



Task Farm Model. Performance Evaluation.



Parallel Implementation on a Metacomputer.



Implementation Issues. Performance Evaluation.



Concluding Remarks. Bibliography.

29. Biomedical Applications Modeling.


Introduction. The Chromosome Reconstruction Problem.



Physical Mapping via Clone Ordering.



PVM Algorithms for Chromosome Reconstruction.



Simulated Annealing and Microcanonical Annealing. Parallel SA and MCA Using PVM.



Heart Rate Variability and Kolmogorov Entropy.



K2 Entropy. Serial Computation of the Correlation Integral.



A Parallel Algorithm for K2 Entropy Computation using PVM. Optimal Scaling Region Determination Algorithm. Experimental Results.



Chromosome Reconstruction. K2 Entropy Computation.



Conclusions. Bibliography.

Appendix A. Glossary.
Index.
show more

About Rajkumar Buyya

RAJKUMAR BUYYA is a researcher at the School of Computer Science and Software Engineering, Monash University, Melbourne, Australia. He was guest editor for the Special Issue on High Performance Computing on Clusters, Parallel and Distributed Computing Practices Journal, and co-author of Mastering C++ and Microprocessor x86 Programming. He is a speaker in the IEEE Computer Society Chapter Tutorials Program and chairman of the IEEE Computer Society Task Force on Cluster Computing.
show more

Rating details

11 ratings
4 out of 5 stars
5 36% (4)
4 27% (3)
3 36% (4)
2 0% (0)
1 0% (0)
Book ratings by Goodreads
Goodreads is the world's largest site for readers with over 50 million reviews. We're featuring millions of their reader ratings on our book pages to help you find your new favourite book. Close X