Parallel Architectures

Available kind of materials:

• Document in HTML format
• Document in (Gzipped) Postscript
• Transparencies in (Gzipped) Postscript

Overview

1. Introduction MIMD architectures
   Document (12 pages Gzipped Postscript 37k) by Péter Kacsuk
   Transparencies (9 slides Gzipped Postscript 22k, Zipped Winword 20k) by Zsolt Németh
   Abstract (HTML)
   1. Architectural concepts
   2. Problems of scalable computers
   3. Main design issues of scalable MIMD computers
2. Introduction to IFP architectures by Péter Broczkó
   Document (21 pages Gzippped Postscript 83k)
   Transparencies (27 slides Gzipped Postscript 60k, Zipped WinWord 62k)
   Abstract (HTML)
   1. Overview of IFP processors
   2. Constraints on parallel instruction execution
   3. The speed-up potential of functional parallelism
3. Superscalar architectures by Dezsô Sima
   Document (Part I. of the document, 27 pages Gzipped Postscript 83k, Zipped WinWord 74k)
   (Part II. of the document, 43 pages Gzipped Postscript 122k, Zipped WinWord 121k)
   (Part III. of the document, 56 pages Gzipped Postscript 172k, Zipped WinWord 134k)
   Transparencies (Zipped Winword 258k)
   1. Introduction
   2. Parallel decoding
   3. Superscalar instruction issue
   4. Shelving
   5. Register renaming
   6. Parallel execution
   7. Preserving the sequential consistency of instruction execution
   8. Preserving the sequential consistency of exception processing
   9. Implementation of superscalar CISC-processors using a superscalar RISC core
   10. Case studies of superscalar processors
4. Multi-threaded architectures
   Document (40 pages Gzipped Postscript 117k) by Péter Kacsuk
   Transparencies (38 slides Gzipped Postscript 115k, Zipped Winword) by Zsolt Németh
   Abstract (HTML)
   1. Introduction
   2. Computational models
   3. Neumann based multi-threaded architectures
   4. Dataflow architectures
   5. Summary
5. Shared memory architectures
   Document (105 pages Gzipped Postscript 333k) by Péter Kacsuk
   Transparencies (72 slides Zipped Winword 22k, Zipped Winword 129k, Zipped Winword 70k) by Etelka Kovács
   1. Introduction
   2. Dynamic interconnection networks
   3. Cache coherence
   4. Synchronization and event ordering in multiprocessors
   5. Uniform Memory Access (UMA) Machines
   6. Non-Uniform Memory Access (NUMA) Machines
   7. Cache-Coherent Non-Uniform Memory Access (CC-NUMA) Machines
   8. Cache-Only Memory Architectures (COMA) Machines
   9. Summary
6. Distributed memory architectures
   Document (70 pages Gzipped Postscript 208k) by Péter Kacsuk
   Transparencies (72 slides Gzipped Postscript 184k, Zipped Winword 382k) by Gyula Fehér
   1. Introduction
   2. Direct interconnection networks
   3. Fine-grain systems
   4. Medium-grain systems
   5. Coarse-grain multicomputers
   6. Summary
7. Case examples:
   • Occam by Péter Broczkó
     Document (112 pages Gzipped Postscript 189k, Zipped WinWord 75k)
     Transparencies (41 pages Gzipped Postscript 59k, Zipped WinWord 23k)
     Abstract (HTML)
     1. Introduction
     2. The conception of communicating sequential processes (CSP)
     3. The programming language OCCAM, based on CSP
     4. Practical works
     5. Appendix
   • Transputers by Péter Broczkó
     Document (35 pages Gzipped Postscript 52k, Zipped WinWord 41k)
     Transparencies (22 pages Gzipped Postscript 36k, Zipped WinWord 18k)
     Abstract (HTML)
     1. General princips of Transputers
     2. Advanced features of Transputers (T9000)
     3. Appendix: Transputer characteristics