Dear Filip,
Thanks for your reaction and sorry for my late reply (have been at
some conference).
On Aug 23, 2008, at 8:10 PM, Filip Konvička wrote:
Actually I experimented with parallel search a few years ago, but I
concluded that it is unreliable (at least on Windows) and so I
abandoned it.
With my first mail on this subject I sent a brief example which works
fine on Mozart 1.3.2 (with 2 processors on my MBP and L=24 the
parallel search actually took less than halve of the time than the
same program with only a single process). Also, I tried a similar
example on a cluster and it worked fine too.
So, it appears to me that parallel search worked fine on UNIX on
1.3.2. But there is now a problem with the new version.
The old Parallel Search engine sits on top of the DP framework, and it
might be that it is not too much code. So maybe you'll be able to
take a
look at it yourself (comparing calls to the 'old api' with the 'new
api'
docs and the like).
I can certainly try and fix the parallel search myself (and I
actually may :), after all this is an open-source project. However, I
never did any distributed programming so far (besides using parallel
search). Therefore I asked whether someone could show by an example
how the distributed programming model changed, to give me a start.
Thank you!
Best
Torsten
--
Torsten Anders
Interdisciplinary Centre for Computer Music Research
University of Plymouth
Office: +44-1752-586219
Private: +44-1752-558917
http://strasheela.sourceforge.net
http://www.torsten-anders.de
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
