Peter A a écrit : > Hello, > > I'm working on a benchmark tool for linear algebra libraries in Java that > measures runtime performance and stability. I have run the 2.0 library > thought it and posted the results online at: > > http://code.google.com/p/java-matrix-benchmark/ > > I think that the developers of this library might find those results > interesting. In addition, feedback on the benchmark itself and how I'm > evaluating commons math would also be appreciated.
Thanks for this work. Up to now we had almost no reliable comparison tools. This will help us improve our library. > > I believe that there are some issues with the SVD implementation in commons > math 2.0. Another person, Holger Arndt, is also working on a benchmark > himself and noticed similar problems. I would post a link to his recent > results but I don't believe he has them up in a public place yet. I have > yet to test the 2.1a code from SVN. We know about the issues in SVD and are working on them. The SVD implementation was new in 2.0 and even if several issues have been fixed since 2.0, some are still pending, even in the svn repository (see for example <http://issues.apache.org/jira/browse/MATH-333>). So we can't even think about performance for this specific operation now. > > - Peter Abeles > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
