Which sparse matrix package?

[mmanns at gmx.net]

On Fri, 19 Dec 2008 08:33:28 +1000
"James Mills" <prologic at shortcircuit.net.au> wrote:

> > The dict that I tried out is of the type:
> >
> > {(1,2,3): "2323", (1,2,545): "2324234", ... }
> >
> > It is too slow for my application when it grows. One slicing
> > operation with list comprehensions takes about 1/2 s on my computer
> > for 1E6 elements.
> 
> Let me get this straight.
> It's taking 0.5s to slice your matrix
> of 1E7 (10000000.0 rows/columns)

My benchmark is as follows:

1) Each of the numbers in the 3-tuple is in the range [0, 1E7).
2) There are 1 000 000 elements in the dict (randomly distributed).
3) The content string is a random number in the range [0, 1E10) that is
casted into a string.
4) Measure the time that retrieving all elements in a 10000x100x10 cube
requires.

> Are you mad ? This is TEN Millions and you
> required it faster than 0.5s ?

Think about how often a spreadsheet re-calculates cells.
Furthermore, people tend to copy&paste cells.

I had some small example (Wagner-Whitin algorithm demo)
with hundreds of slicing operations on one screen. I believe that
no-one wants to wait for 1/2 s times 500 == 250 s each time the
spreadsheet is updated.


Martin