CS
Honours course 7933: Distributed and High-Performance Computing
Course Outline and Lecture Notes
Copies of slides presented at the lectures will be provided below
in PDF and PostScript (PS) format.
-
Introduction to Distributed and High-Performance Computing:
What is DHPC; Milestones in DHPC history; Some DHPC applications.
PDF |
PS
-
High-Performance Computing architectures:
Overview of the major classes of HPC architectures and their evolution.
PDF |
PS
-
Parallel programming models and performance analysis:
Parameterisation, modelling, performance analysis, efficiency, and
benchmarking of DHPC systems.
PDF |
PS
-
Programming parallel computers:
Overview of parallel programming,
parallel languages, parallelizing compilers,
message passing and data parallel programming models.
PDF |
PS
-
Message passing programming and MPI:
Uses; historical background and use on MIMD machines;
current implementations;
programming using the Message Passing Interface (MPI).
PDF |
PS
-
Data parallel programming and HPF
Data parallel programming paradigm; historical background and use
on SIMD machines; array syntax;
Fortran 90 and High Performance Fortran (HPF).
PDF |
PS
-
Shared memory programming, threads and OpenMP
Use of shared memory machines; threads; mutual exclusion;
locks, semaphores and monitors; parallel Java;
programming using OpenMP.
PDF |
PS
-
Case Study - Monte Carlo Simulation of the Ising Model:
Case study from computational physics; different approaches to
parallelism; regular vs irregular problems;
techniques and tricks for parallel implementation in MPI, HPF,
OpenMP and threads.
PDF |
PS
-
High-performance distributed computing:
Distributed and concurrent computing on loosely coupled distributed systems;
motivation and applications; transparency and metacomputing;
networks and clusters of workstations; cluster management systems.
PDF |
PS
-
Distributed computing middleware:
Middleware; RPC; DCE; CORBA; Java RMI.
-
Time and ordering in distributed systems:
The problem of time and ordering in distributed systems;
vector time; fault tolerance; Byzantine Generals problem.
PDF |
PPT
-
Grid computing:
The Grid;
Grid computing (metacomputing over wide-area networks);
grid computing environments (Globus, Legion, DISCWorld, ...);
Internet computing (SETI@Home, etc).
Slides for the 2000 DHPC course
are also available.
Distributed and High-Performance Computing course,
Paul Coddington
(paulc@cs.adelaide.edu.au),
July - October 2000.