I have a generalized array module (in C) that would store large
arrays/matrices of floats quite a bit more efficiently than Python
lists. I have also written a generalized N-dimensional array module
(in Python) that you can use with the array module or with ordinary
lists to implement matrix ops. The latter is really still under
development. I should also warn you that "the usual ops on matrices"
is a very ambiguous statement, if you need numerical operations like
inversion you're out of luck. I wouldn't dream of writing numerical
code in Python; nor would I dream of writing it in C; I would just
link to an existing numerical library...
--Guido van Rossum, CWI, Amsterdam <Guido.van.Rossum@cwi.nl>