[Numpy-tickets] [NumPy] #594: Major slowdown when mixing array with Python's list

NumPy numpy-tickets@scipy....
Sat Dec 15 23:03:50 CST 2007


#594: Major slowdown when mixing array with Python's list
------------------------+---------------------------------------------------
 Reporter:  Boris       |        Owner:  somebody
     Type:  defect      |       Status:  new     
 Priority:  normal      |    Milestone:  1.0.5   
Component:  numpy.core  |      Version:          
 Severity:  normal      |   Resolution:          
 Keywords:              |  
------------------------+---------------------------------------------------
Comment (by cdavid):

 I don't think it is a bug per se. When you index with a list, you need to
 convert it to an array or something internally (I don't know how this
 works exactly inside numpy, but for each item of the list, at some point,
 you will have to check it is an integer, convert it to an integer at the
 machine level: this is really slow).

 Why do you need to create so many arrays ? Can't you preallocate them, for
 example (specially if they are all of the same size) ?

-- 
Ticket URL: <http://scipy.org/scipy/numpy/ticket/594#comment:3>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.


More information about the Numpy-tickets mailing list