Re: Parallel computing and JLink
- To: mathgroup at smc.vnet.net
- Subject: [mg94433] Re: Parallel computing and JLink
- From: mark mcclure <mcmcclur at unca.edu>
- Date: Fri, 12 Dec 2008 06:59:06 -0500 (EST)
- References: <ghg9bb$ho5$1@smc.vnet.net> <ghlm9k$jgt$1@smc.vnet.net>
On Dec 11, 3:44 am, Jens-Peer Kuska <ku... at informatik.uni-leipzig.de> wrote: > When I have to switch to Java for the computation I lose > the "wonderful Mathematica language" and than I can also > use C/C++, FORTRAN or anything else. > > And if it should be fast, I mean really fast one has to > use C/C++ anyway. Here's a neat picture: http://facstaff.unca.edu/mcmcclur/g2space-22.png The picture shows a 2D slice of the 4D parameter space for the complex dynamics of Weierstrass elliptic functions. This a topic of current research in complex dynamics that I found myself collaborating on via email with a colleague. To generate such an image directly in Mathematica is just *way* too slow, since evaluation of a Weierstrass elliptic function is *much* more expensive than evaluation of a quadratic. JLink was the perfect tool to use to speed up the computation. Not only is JLink easier to use than the more general MathLink, but cross-platform compatibility was not a problem -- I'm a mac; she's a PC. Sure, an alternate language might be a few times faster, but the Java speedup was more than sufficient to generate some publication quality pictures. Of course, MathLink and C are great as well and I personally appreciate the great work you've done with them. Thanks, Mark McClure
