JRuby and linear equations

[Axel Etzold]

Dear all,

can you recommend a good and fast way of calculating the solutions
of linear equations in JRuby, including singular value decompositions of
matrices (http://en.wikipedia.org/wiki/Singular_value_decomposition) ?

Thank you very much!

Best regards,

Axel
--
GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[BjoernW]

There are two packages for Java:

http://acs.lbl.gov/~hoschek/colt/index.html
http://math.nist.gov/javanumerics/jama/

I'd go for Colt. If it's good enough for CERN it should be good enough
for everyone :)

2009/10/18 Axel Etzold <AEtzold@...>:

> Dear all,
>
> can you recommend a good and fast way of calculating the solutions
> of linear equations in JRuby, including singular value decompositions of
> matrices (http://en.wikipedia.org/wiki/Singular_value_decomposition) ?
>
> Thank you very much!
>
> Best regards,
>
> Axel
> --
> GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
> Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01
>
> ---------------------------------------------------------------------
> To unsubscribe from this list, please visit:
>
>    http://xircles.codehaus.org/manage_email
>
>
>

--
Viele Grüße / Best regards, Björn Wilmsmann
Geschäftsführer / CEO

----------------------
MetaSieve GmbH

Universitätsstr. 142
D-44799 Bochum
Germany

Phone: [phone]
Mobile: [phone]
Fax: [phone]

E-mail: bjoern@...
http://www.metasieve.com/

Amtsgericht Bochum, HRB 12288
----------------------

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[john casu-3]

Björn Wilmsmann wrote:
> There are two packages for Java:
>
> http://acs.lbl.gov/~hoschek/colt/index.html
> http://math.nist.gov/javanumerics/jama/
>
> I'd go for Colt. If it's good enough for CERN it should be good enough
> for everyone :)
>


latest change on Colt == Sep 10, 2004.
last change on Jama == July 13, 2005

no-one doing serious numerical work, uses Java.  almost by definition.
and both of these packages, seem to be fragile/underdeveloped.

Since FFI works with Jruby, I'd be inclined to roll your own limited
linear algebra toolkit.  take a look at linalg.rubyforge.org to see how
this was done with MRI 1.8

> 2009/10/18 Axel Etzold <AEtzold@...>:
>> Dear all,
>>
>> can you recommend a good and fast way of calculating the solutions
>> of linear equations in JRuby, including singular value decompositions of
>> matrices (http://en.wikipedia.org/wiki/Singular_value_decomposition) ?
>>
>> Thank you very much!
>>
>> Best regards,
>>
>> Axel
>> --
>> GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT!
>> Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01
>>
>> ---------------------------------------------------------------------
>> To unsubscribe from this list, please visit:
>>
>>    http://xircles.codehaus.org/manage_email
>>
>>
>>
>
>
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[BjoernW]

2009/10/19 john casu <john@...>:

> latest change on Colt == Sep 10, 2004.
> last change on Jama == July 13, 2005
>
> no-one doing serious numerical work, uses Java.  almost by definition. and
> both of these packages, seem to be fragile/underdeveloped.

Well, the basics of linear algebra like matrix operations are pretty
stable. These are more or less standard operations like + or - so
there's no reason why there should be ongoing development in this area
if the existing software works correctly, which it probably does if
CERN uses it.

--
Viele Grüße / Best regards,
Björn Wilmsmann

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email