Ok I finally found the error. I should have used cpdef for the _lmul_
function. Now sparse*scalar multiplication runtime is finally no
longer depending on the matrix size, only on the number of nonzero
enties. It's ready for review at http://trac.sagemath.org/sage_trac/ticket/10568

BTW. If you read this william, this also undo's the major slowdown of
the modular symbols initialization code we discussed about earlier.

On Jan 7, 6:31 pm, koffie <[email protected]> wrote:
> I tried to speed up scalar multiplication of sparse matrices.  I did
> this using _lmul_ as explained 
> inhttp://www.sagemath.org/doc/reference/coercion.html
>
> I alredy made a patch giving sparse matrix objects a function _lmul_
> but somehow it doesn't seem to get called.
>
> Seehttp://trac.sagemath.org/sage_trac/ticket/10568for the ticket I
> opened it has a patch so you can exactly see what I did.

-- 
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to