Hi Demian-
PDL::LinearAlgebra is bindings to existing "gold standard"
BLAS and LAPACK linear algebra routines. I think you'll be
ok with numerics from there.
Regarding linear algebra for PDL, that is one of the areas
we've long desired improvement. I've adopted PDL::LinearAlgebra
for maintenance and in the hopes of improving the library
detection and build. However, the GNU Scientific Library has
a large number of computational routines including linear
algebra that could use PDL bindings.
Cool to hear about your iperl notebook work. Would you be
interested in making an intro PDL tutorial for it? PDL-2.008
is planned for this month. It would be nice to have some
flash and coolness with the announcement. :-)
--Chris
On 2/4/2015 22:03, Demian Riccardi wrote:
Hello everyone,
I’m working on a few iperl notebooks for a publication that I’m trying
to push through. I want to show an analysis of protein fluctuations
and correlated motion using a gaussian network model (GNM). GNM is
easy to implement, you just need to calculate a Kirchoff matrix and
then compute the pseudo inverse (first eigenvalue is zero). See the
following:
https://github.com/demianriccardi/HackaMol/blob/master/examples/Notebooks/HackaMol-GNM.ipynb
I use PDL::LinearAlgebra to calculate the pseudo inverse, but I felt a
little squeamish when I downloaded it and looked around at the test
suite. Is PDL::LinearAlgebra my best option, or are there other
recommended ways of going.
Thanks!
Demian
p.s. Zaki’s iperl notebook has changed my life almost as much as moose
has.
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
