[Numpy-discussion] Good way to develop numpy as popular choice!

Travis Oliphant travis@continuum...
Fri Jun 22 10:13:12 CDT 2012


> 
> 
> -Travis
> 
> 
> However, what is the timing/memory cost of converting a large numpy array that already exists into python list of lists?  If all my processing before the munkres step is using NumPy, converting it into python lists has a cost.  Also, your timings indicate only ~2x slowdown, while the timing tests done by eat show an order-of-magnitude difference.  I suspect there is great room for improvement before even starting to worry about the array access issues.
> 

If you are also doing scalar math with the results returned from NumPy array element access, then a.item(i,j) will be faster because it returns a Python object which will use it's scalar math instead of re-using vectorized math operations which a[i,j] will do. 

I haven't looked at the code yet, just re-emphasizing known issues with trying to use NumPy as an arbitrary container of "elements" rather than a container of bytes that you do vectorized operations on. 

-Travis

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/numpy-discussion/attachments/20120622/12d3032d/attachment.html 


More information about the NumPy-Discussion mailing list