[SciPy-dev] Tri-diagonal LAPACK Routines - Shall I interface them?

David Warde-Farley dwf@cs.toronto....
Tue Dec 1 20:06:40 CST 2009


On 1-Dec-09, at 8:34 PM, Simon Clift wrote:

> Is there a broader reason for not completing the interface?

Nope. I think they were just wrapped on an as-needed basis. This came  
up a little while ago w.r.t. general banded systems as well:

	http://mail.scipy.org/pipermail/scipy-user/2009-October/023083.html

> Would anyone object if I went ahead and patched scipy/linalg/ 
> generic_flapack.pyf to allow
> the call?


Doubtful. :) Although if you do, it might be a good idea to add a  
function to scipy.linalg with a comprehensible name (like  
solve_tridiagonal to go along with solve_banded [which actually does  
wrap dgbsv, my post there is wrong in that regard]).

David



More information about the SciPy-Dev mailing list