Parallel optimization algorithms and implementations
Robert G. Brown
rgb at phy.duke.edu
Wed Mar 12 06:30:11 PST 2003
On Tue, 11 Mar 2003 Brian.M.Ironside at frb.gov wrote:
> Hello All,
> I've been searching for resources on the topic of parallel algorithms
> for function minimization, and existing implementations of such algorithms.
> So far, I've come up with very little. I was hoping someone on this mailing
> list might be able to point me in the right direction. From what little
> research I've done into serial minimization algorithms, there certainly
> seems to be room in a number of algorithms for speed-up through parallel
> computation, and I'd love to know if anyone was done anything with this.
You mean like books on parallel programming? Yes, there are several,
although I wouldn't quite say there are a lot. Of course, I don't
really know how many there are as I haven't done a really systematic
search for them, just encountered a few along the way, so to speak.
Real Computer Scientists have been messing with parallel algorithms, MP
systems, communications models, and all of that for many years, and I'm
certain that useful and interesting research continues in the field
A few "standard text" type references:
G. S. Amalsi and A. Gottlieb, Highly Parallel Computing (2nd edition),
I. Foster, Designing and Building Parallel Programs, Addison-Wesley,
1995. Also see the online version of the book at Argonne National Labs,
V. Kumar, A. Grama, A. Gupta, and G. Karypis, Introduction to Parallel
Computing, Design and Analysis of Algorithms, Benjamin/Cummings, 1994.
Probably some list person can provide some post mid-90's texts, and an
amazon or google search would likely turn up some as well. You might
start with Foster's book, since it is online and is a partial model for
what I'm trying to do with my own online book, except that I haven't yet
finished a deal with AW or any other publisher because I literally do
not have time.
Overall I'd rather buy a book from somebody that provides it online for
free than from somebody who doesn't, in a paradoxical sort of way. A
paper copy is a specific, not horribly portable resource that one can
read deeply and study "at home", the online copy means that no matter
where you are, you can access chapter X and review something you learned
from your deep reading.
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb at phy.duke.edu
More information about the Beowulf