Go up to
Top
Go forward to
Exercises
Contents
Introduction
(
PostScript Slides
)
A cross-section of the course.
Parallel Program Design
(
PostScript Slides
)
A methodological approach to parallel program design.
Performance of Parallel Programs
(
PostScript Slides
)
Measures and models of parallel program performance.
Vector Programming
(
PostScript Slides
)
Pipelining, vector operations, automatic vectorization.
Convex SPP System Overview (
PostScript Slides
)
Hardware, operating system, development tools; the automatically parallelizing compiler.
Shared Memory Programming
(
PostScript Slides
)
Algorithms for matrix multiplication: row/column oriented algorithm, block-oriented algorithm; shared memory synchronization: locks, semaphores, barriers.
MPI Message Passing Interface (
Chapter from Foster's Book
) (
Example Programs
)
A tutorial on MPI programming.
Distributed Memory Programming
(
PostScript Slides
)
Distributed memory algorithms for matrix multiplication: SIMD algorithms, row-oriented algorithm; embedding into hypercube; block algorithm.
Performance Analysis
(
PostScript Slides
)
Data collection, transformation, and visualization; profiles, counters, traces; visualization tools.
Multiprocessor Architectures
(
PostScript Slides
)
Shared memory architectures; cache consistency; distributed memory architectures, interconnection topologies, message routing, broadcasting; distributed shared memory architectures, NUMA, COMA; commercial systems.
Basic Methods
Scheduling algorithms; deadlock detection and avoidance; consumer/producer problem; reader/writer problem; distributed random number generation.
Maintained by:
Wolfgang Schreiner
Last Modification: December 23, 1997
