[SciPy-dev] I'm doing some more experimenting

[rob]

It turns out that what is happening is that blitz() is creating some
overhead which makes my memory consumption higher.  That is why a
100x100x100 wont work on my laptop with blitz, while is works with plain
numpy.  I dropped it down to 80x80x80, and now everything is Ok.

I will try to get an exact amount for the extra memory overhead using
top for 80x80x80.  Right now top says 188Meg for the blitz version.  I
will have to make another file to use plain python.  It is definitely
true that the exact same blitz libraries work with the larger arrays,
until I hit my swap limit.  Rob.



-- 
-----------------------------
The Numeric Python EM Project

www.pythonemproject.com