NA Digest Friday, April 24, 1987 Volume 87 : Issue 37

This weeks Editor: Gene Golub

Today's Topics:


Date: Thu, 23 Apr 87 08:14:12 cst
From: Suter Bruce <suter%uab.csnet@RELAY.CS.NET>


Please send the following message out over NAnet.




to participate in research that will lead to the development of the next
generation of partial differential equation solvers using parallel processing
technology. This endeavor will take place in the Department of Computer and
Information Sciences at the University of Alabama at Birmingham, and will
include interactions with several researchers from both Livermore and
Argonne National Laboratories. This project will utilize the department's
30-processor Sequent Balance 21000 computer ( the largest number of processors
commercially available in a tightly coupled shared memory architecture ) as well
as the new Cray XMP supercomputer being acquired by the Alabama Supercomputer
Authority and accessible January 1988.

The minimum student background should include the following:

Mathematics Computer Science
------------- ------------------

Calculus Data Structures
Differential Equations Computer Organization / Architecture
Linear Algebra Operating Systems
Numerical Analysis

Since this research will involve topics from both Mathematics and Computer
Science, perspective students should feel comfortable in both areas. Student
Support is available. U.S. Citizenship is required. For more information

Prof. Bruce Suter
Dept. of Computer and Information Science
University of Alabama at Birmingham
University Station
Birmingham, AL 35294

e-mail: suter@uab (csnet)
suter%uab.csnet@csnet-relay (arpanet)


Date: Fri, 24 Apr 87 08:48:03 EDT
From: (mcbryan)
Subject: Re: Where is Oliver McBryan

Oliver McBryan was reportedly seen near the summit of Mt. McKinley looking for
an outlet for his modem. He was muttering something about escaping from
conference proceeding editors and indicated he felt he had found the right
location. He was hoping not to find an outlet under the snow, but later
reports indicate that he was seen peering into a screen with several
vultures hovering overhead.

He indicated that his attempt to develop an SIMD parallel paper generation
approach had failed when one paper required a period at the same moment
that another required a comma. He also hinted at a problem with send/receive
protocols - something about too many send requests per receive I think it was
- and said that having to always write a power-of-two papers at once was
an unnecessary restriction.

He is now developing parallel techniques for the Brooks Range, involving
climbing with one foot on each of two mountains. Since the mountains are of
equal height he expects the SIMD approach to work. He was worried about
some long-range communication overheads however. He closed with the
promise that if the new techniques are successfull he will submit parallel
grant proposals to all funding agencies.


End of NA Digest