[computer-go] Language War!

[Jim O'Flaherty, Jr.]

Peter,

I have not used it, yet.  I have dl'ed it and played with it a bit.  I have not had the opportunity to play at that level of performance tuning and memory efficiency yet.  When I do get the chance, I will definitely be checking out how it impacts my Java implementations before attempting a move to something like highly optimized C or assembly.  At this point, my time-value focus is to reduce total time spent coding to achieve effects and then purchase horse-power (much much cheaper than my personal time) to produce results.

BTW, given you use standard java.util.Collection interfaces, then using Javolution implementations ought to be close to transparent swap-ins for whatever you are currently using.  While that is not labor free, it certainly ought to significantly reduce the amount of work you have to do to replace all your instances of "new ArrayList()" with "new FastList()" (or whatever his particular ArrayList replacement is).  And given you are using some sort of source code version control, you can contain the impact to an experimental throw-away branch.  If the branch happens to work out and you want to fully depend upon the Javolution library, you can then merge the changes into your main branch.  When I do finally get personal time and priority to play again in the area, this is how I will be approaching it.


Jim


----- Original Message ----
From: Peter Drake <drake at lclark.edu>
To: computer-go <computer-go at computer-go.org>
Sent: Monday, July 31, 2006 4:20:44 PM
Subject: Re: [computer-go] Language War!


On Jul 31, 2006, at 1:32 PM, Jim O'Flaherty, Jr. wrote:

Peter,

I have a similar impression about size, although I cannot verify it right this second.

Given where you are, have you checked Javolution?
http://javolution.org/

This guy is a Java performance/memory footprint nut-case.  If you look at his posting on http://www.javalobby.org, you will see he stays right on top of performance and footprint metrics around the core collection class replacements he has developed.

Given your framework is suitably designed, it might take you a very small window of time to try out some of his "replacements" and see what kind of improvements you experience, aasuming you have some baselines against which to measure the resulting differences.




Interesting.  Have you (or anyone else) tried this out?  I'm wary of doing anything dependent upon a (relatively) new third-party tool.

Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/




_______________________________________________
computer-go mailing list
computer-go at computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://computer-go.org/pipermail/computer-go/attachments/20060731/b9f2a662/attachment.htm