Seeking SVD code

[Tristan Miller]

Greetings.

Can someone point me towards some portable source code which will
efficiently perform singular value decompositions of large matrices? I am
looking for something freely available, preferably in the public domain or
under a Free or Open Source licence such as the GPL. Ideally the program
will be in C, though any languages compilable under Unix/Linux and MS
Windows without proprietary tools such as Matlab are OK.

Regards,
Tristan

--
_
_V.-o Tristan Miller [en,(fr,de,ia)] >< Space is limited
/ |`-' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= <> In a haiku, so it's hard
(7_\\ http://www.nothingisreal.com/ >< To finish what you

[Tim Heaney]

Tristan Miller <psychonaut@nothingisreal.com> writes: Can someone point me towards some portable source code which will efficiently perform singular value decompositions of large matrices? I am looking for something freely available, preferably in the public domain or under a Free or Open Source licence such as the GPL. Ideally the program will be in C, though any languages compilable under Unix/Linux and MS Windows without proprietary tools such as Matlab are OK.

How about LAPACK?

http://www.netlib.org/lapack/

I hope this helps,

Tim

[Roland Van Praag]

You can try JAMA with TNT
It is C++ and Free

http://math.nist.gov/tnt/download.html

Regards,
Roland


"Tristan Miller" <psychonaut@nothingisreal.com> wrote in message
news:1323239.iQDmn99nfD@ID-187157.news.dfncis.de... Greetings. Can someone point me towards some portable source code which will efficiently perform singular value decompositions of large matrices? I am looking for something freely available, preferably in the public domain or under a Free or Open Source licence such as the GPL. Ideally the program will be in C, though any languages compilable under Unix/Linux and MS Windows without proprietary tools such as Matlab are OK. Regards, Tristan -- _ _V.-o Tristan Miller [en,(fr,de,ia)] >< Space is limited / |`-' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= <> In a haiku, so it's hard (7_\\ http://www.nothingisreal.com/ >< To finish what you

[Roger Schlafly]

"Roland Van Praag" <roland@numeca.be> wrote You can try JAMA with TNT It is C++ and Free http://math.nist.gov/tnt/download.html

Does it work on any common compilers? It uses a lot of tricky
C++ templates, and compiler support for those features tends
to be spotty.

[Roland Van Praag]

I needed to do the followings before getting it running
on several compilers
here as example: tnt_array1D

1) remove namespace that is not supported by all compilers
replaced by adding TNT_ to the class name
2) remove explicit that is not supported by all compiler
replace TNT_Array1D by TNT_Array1D<T> within the copy function
3) In the class declaration:
replace all instances of TNT_Array1D by TNT_Array1D<T>
exept for constructors and destructors

Regards,
Roland


"Roger Schlafly" <rogersc@mindspring.com> wrote in message
news:n_Tpb.4871$tL7.1446386826@twister1.starband.net... "Roland Van Praag" <roland@numeca.be> wrote You can try JAMA with TNT It is C++ and Free http://math.nist.gov/tnt/download.html Does it work on any common compilers? It uses a lot of tricky C++ templates, and compiler support for those features tends to be spotty.

[bv]

Tristan Miller wrote: Can someone point me towards some portable source code which will efficiently perform singular value decompositions of large matrices? I am looking for something freely available, preferably in the public domain or under a Free or Open Source licence such as the GPL. Ideally the program will be in C, though any languages compilable under Unix/Linux and MS Windows without proprietary tools such as Matlab are OK.

The "arpack" library in http://netlib.org is a good place to start.

--
Dr.B.Voh
------------------------------------------------------
Applied Algorithms http://sdynamix.com