[SciPy-user] Optimization in numexpr for unaligned arrays

[Francesc Altet]

Hi,

I've filed a patch with ticket #400 of SciPy trac. I've tried to assign
it to David, but it seems that I've not privileges to do this.

Anyway, I'm puzzled about why, when working with unaligned arrays, a
plain "a.copy()" in Python space can be faster (between a 30% and a 70%)
than a 'PyArray_SimpleNewFromDescr' call in C space.

Any hints?

-- 
Francesc Altet    |  Be careful about using the following code --
Carabos Coop. V.  |  I've only proven that it works, 
www.carabos.com   |  I haven't tested it. -- Donald Knuth