Hi Rob,

On 14 Feb., 23:42, Rob Beezer <[email protected]> wrote:
> Hi Simon,
>
> You know the .new_matrix() method, as well?

Of course. That's where all the overhead that I mentioned is created,
directly or indirectly. So, the aim is to replace the .new_matrix()
code by something faster. I partially succeeded (see #8611 [merged]
and #10763 [needs review]), but one bottle neck remains.

>  It does not currently
> make an "empty" matrix, but defaults to a zero matrix.

Sure. And the question is whether there really is no faster way to
create a zero matrix, than to *copy* a zero matrix (this is what is
currently done in the dense case). Perhaps it is naive, but I thought
that writing should be faster than reading and writing.

Best regards,
Simon

-- 
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