Re: Parallel computing and JLink
- To: mathgroup at smc.vnet.net
- Subject: [mg94220] Re: Parallel computing and JLink
- From: Syd Geraghty <sydgeraghty at me.com>
- Date: Mon, 8 Dec 2008 06:22:21 -0500 (EST)
- References: <200812071042.FAA18184@smc.vnet.net>
Mark, I have downloaded your > You can download a gzipped archive of a folder with the correct > hierarchy here: > http://facstaff.unca.edu/mcmcclur/Mathematica/ParallelJLink/ParallelJLink.tar.gz > . I really appreciate your efforts here and have been delighting in exploring (and trying to understand) what you have done. Everything worked fine and the specified number of kernels could be tracked using SystemInformation[] in Mathematica and the Mac Activity Monitor. I want to report that after setting up 4 kernels (actually just evaluating your QuadraticIterationInParallel.nb which includes this) I reran the MathematicaMark7 benchmarks to compare with the results I got immediately after upgrading a few weeks ago. Machine Name: 4-node homogeneous cluster System: MacOSX-x86-64 Date: December 7, 2008 Mathematica Version: 7.0.0 Benchmark Result: 2.30 Compared to: Machine Name: localhost System: Mac OS X x86 (64-bit) Date: November 21, 2008 Mathematica Version: 7.0.0 Benchmark Result: 1.92 As you can see the benchmark result improved from 1.92 to 2.30. Also note the Machine Name change to "4-node homogeneous cluster"! (The elapsed time seemed to be longer but I was running with lots of other processes around). That was very impressive and puzzling. I had assumed that a default environment would have been set up by the Mathematica 7 distribution that would take care of optimizing the Mathematica system based on SystemInformation being discovered during installation. This is evidently not the case. Question to you and MathGroup: Where do I find out how to optimize and use the multiprocessor capabilities of my Mac? I have had a look through the Documentation Center and have concluded it could take a very long time to find the answer. I have no wish to be negative or stoke the "Documentation wars" again but it is clear that without following up on your post I would not have easily got to discovering how to use Mathematica in such an optimized way as you do. Cheers ... Syd Syd Geraghty B.Sc, M.Sc. sydgeraghty at mac.com Mathematica 7.0.0 for Mac OS X x86 (64 - bit) (21st November, 2008) MacOS X V 10.5.4 MacBook Pro 2.33 Ghz Intel Core 2 Duo 2GB RAM On Dec 7, 2008, at 2:42 AM, mark mcclure wrote: > Jon Harrop makes an interesting point in this post: > http://forums.wolfram.com/mathgroup/archive/2008/Dec/msg00102.html > > If we need to optimize code so much that we are willing to > parallelize it (which can be quite tricky), then why wouldn't > we use optimized C or Fortran code? In fact, this very logic > persuaded me against purchasing Grid Mathematica, even though > I had potential access to a cluster. Now that parallel tools > are incorporated into every copy of Mathematica, however, why > not use it. In fact, it's quite likely that people will come > up with many novel ideas. > > Here's one proposal: use Mathematica together with JLink to > coordinate Java programs running in parallel. (Of course, > mathlink and C would work as well.) In this scenario, > Mathematica becomes much more than just a master program > coordinating a few Java slaves, because any data generated by > the Java programs becomes immediately available to Mathematica. > A natural application is Mandelbrot or Julia set generation. > I set up a notebook/webpage illustrating this. You can see it > here: > http://facstaff.unca.edu/mcmcclur/Mathematica/ParallelJLink/ > > Since there are a few difficulties and traps, I set up as > carefully as I could and included a fair amount of explanation. > > Mark McClure >
- References:
- Parallel computing and JLink
- From: mark mcclure <mcmcclur@unca.edu>
- Parallel computing and JLink