LinuxJournal has a nifty article on how to build simple clusters with Ruby, based on real world experience from the National Geophysical Data Center. Article here
Clusters with Linux and Ruby
Submitted by riffraff 2004-12-06 General Development 7 Comments
That was an interesting read. Ruby looks quite cool.
I only would like it was made in python, but as simple as it is it can be easily ported to any language. No, not java, or you will end with a cluster slower than a 386 (this was a joke, obviusly)
Thanks for the article. I like ruby. Ruby makes programming fun. I wish more people would give it a try.
Similar to the second comment, I think python is probably the more heftier language – not that that necessarily makes it always the better choice, mind you.
Really interesting article.
BTW if anyone is interested in this theme there is a group I set up at yahoo
for discussion of distributed Ruby.
Ruby is so powerful! Im trying to move all my script work to it, because it has so mutch features, the code is actualy readable. This are nice days for all the developers: everyone can do, what suits’em. From Borland’s IDE to Mono, C, C++, C#, Python, Perl, Ruby, Java, PHP. Just take whats you need, and code. Mniam.
I’ve been using Python for some light numerical programming for a while so that I don’t have to stay with FORTRAN all the time. Is Ruby still much slower than Python? Has there been recent developments on speeds?
if you use numerical libraries such as R for ruby or ruby/idl or ruby/GSL or NArray you should have great performance (just like you would using numarray/numeric or python/gsl).
OTOH I think the ruby basic interpreter is somewhat slower than python’s one, but they’re both way too much slow to be used for numerical stuff that way in any case. A matrix calculation in pure python would appear nearly 1000 timesd slower than fortran withouth numeric/narray.