This directory contains files for Technical Report CSRI-389.

The files in this directory are the following:

1) README       (   2 KB - the ASCII file you are now reading)
2) TR-389.ps    (3185 KB - PostScript)
3) TR-389.ps.gz ( 593 KB - PostScript compressed with the program "gzip")

If you have the UNIX "gunzip" program, get the file TR-389.ps.gz.
Remember to transfer the file in binary mode. After the transfer, 
"gunzip" the file.

If you do not have the UNIX "gunzip" program, get the file TR-389.ps 
in ASCII mode. 

After transferring the file, print it on a PostScript printer.

If you have any questions or comments about this technical report, please
contact glynch@cs.toronto.edu


TITLE:    Parallel Job Scheduling on Heterogeneous Networks of
          Multiprocessor Workstations 

AUTHOR:   Gerard Lynch


ABSTRACT:

Parallel job scheduling has been heavily studied on various
architectures, from vector supercomputers to networks of 
workstations.  A common assumption of this previous work has 
been that all processors in the system have equal processing 
power. We investigate parallel job scheduling on heterogeneous
networks of multiprocessor nodes, where the machines in the 
network can vary in processor speed and node size (number of
processors per node). 

Two dynamic scheduling policies and two adaptive scheduling 
policies are simulated in this framework.  We provide some evidence
that a network's coefficient of variation (standard deviation over
mean) of processor speed or node size is a good predictor of job mean
response time for these algorithms.  As well, processor-speed 
heterogneity is found to have a greater effect on job mean response
time than node-size heterogeneity.  In some cases, the adaptive
algorithms can outperform the more flexible dynamic algorithms 
due to job imbalances created by dynamically reassigning processors.