[Numpy-discussion] array vs. matrix performance

[Robert Cimrman]

Re-hi,

thanks for all the comments. I have re-tried with
X = nm.random.rand( 10000, 3 ) and the times (in seconds) were:

428.588043213 # scipy.dot, array
445.045716047 # numpy.dot, array
519.489458799 # scipy.dot, matrix
513.328601122 # numpy.dot, matrix

The scipy.dot and numpy.dot performs the same for me, as I use ATLAS,
IMHO. David, you are right that for larger data the difference is not so
big but still they are significant. Anyway, I have resolved my problem
with arrays only for the moment.

r.