From help-request at octave dot org Thu Jan 27 01:59:39 2005 Subject: Re: Octave of MP systems From: Javier =?iso-8859-1?Q?Fern=E1ndez?= Baldomero To: Jonathan Greenberg CC: "help at octave dot org" , javier@atc.ugr.es Date: Thu, 27 Jan 2005 08:58:32 +0100 Hi Jonathan, Jonathan Greenberg escribió: > > Has anyone used octave on huge MP systems (# processors > 256)? Is there a I have tried octave in a cluster of 8 computers, and Michael Creel has written a paper on econometrics using octave on 16 computers. Of course, I'm also interested in knowing of the most node-spreaded octave parallel application so far :-) Michael used MPITB for his parallel applications. I keep track of them at http://atc/javier-bin/mpitb (this is mine) http://pareto.uab.es/wp/2005/63705.pdf (this is Michael's) > listing of MP-optimized algorithms anywhere, and a discussion of issues > concerning Octave MP support? > Again, I'm also interested if you find something on that subject. There are the NAS Parallel Benchmarks, and people porting them to optimized MATLAB code, like http://www.ece.osu.edu/~bussb/research/ Some time ago I prepared a "scrapbook" to make up my mind about the announcements and discussions on MPI in the octave mailing lists. In a first section I extracted the download links as I found them mentioned in the posts. In the second section I chronologically ordered the different posts that I could find dealing on parallel Octave, adding a small remark summarizing each thread. I made a poster for the VecPar Conference (also linked in MPITB web above) which used that info. I still keep it, but all remarks are in spanish. Drop me an e-mail and I'll gladly send it to you. Also drop me an e-mail or post again here if you find more information on Octave on MP :-) Last year I was looking for time-consuming octave applications to port to MPITB (in the end it's been Michael who has done the work :-). BTW, I would be glad if somebody helped me choose at least one Chemicals application that I could port to MPITB, for several reasons: - Octave was a ChemEng professor, the fastest when it came to calculations :-) * http://www.octave.org/history.html - There are strong research lines in ChemEng Depts dealing on computer simulation * http://www.octave.org/history.html again Octave helped ChemEng students with reactor design * http://capd.cheme.cmu.edu/newsletters/March_03.PDF for instance, in this link you will hardly find a research line not dealing with time-consuming computations/simulations - The examples of the reactor design book are online :-) * http://www.che.wisc.edu/jbr-group/ http://www.che.wisc.edu/~jbraw/chemreacfun/ If only somebody could help me choose one of those examples to parallelize :-) (one that is time-consuming and can be somehow split in independent subtasks) -javier ------------------------------------------------------------- Octave is freely available under the terms of the GNU GPL. Octave's home on the web: http://www.octave.org How to fund new projects: http://www.octave.org/funding.html Subscription information: http://www.octave.org/archive.html -------------------------------------------------------------