Greedy String Tiling : Help for speedup an implementation of the algorithm

Hi All

I've implemented the Greedy String Tiling algorithm to find similar strings. This method returns a number which means how similar two strings are. A bigger number indicates more similarity (it is not a percentage of similarity). We are using this implementation in an application that generates descriptions for Real Estates. However the performance of the code is low so that to generate descriptions for 1000 properties we have to wait for 5 hours!!

Is there anyone who can help me with the improvement of the performance? The performance improvement will be involved in the implementation strategy rather than optimizing the code. If you are content to help me please let me have your email address so that I can send you the Visual Studio 2010 solution.

P.S. Does anybody have a c# or vb.net implementation of Greedy String Tiling, so that I can compare mine with it?

Cheers

5 hours! for only 1000 descriptions?

You'll find that you have more luck in getting help if you post the troublesome code, properly using [code][/code] tags rather than asking for an email address as most of us would be unwilling to provide our email not to mention it does other readers no good if not posted in the forums.

Yeah, you can attach the project here ("Build Menu" -> "Clean Solution" first to minimize file size) - or just post the code (use the [code][/code] tags). Since string's are immutable in C#, it's recommended to use the StringBuilder class for extensive operations on strings for performance reasons, but even so - 5 hours?!

There must be something else, something in the way it's implemented...