[Numpy-tickets] [NumPy] #595: sum function slow

[NumPy]

#595: sum function slow
--------------------+-------------------------------------------------------
 Reporter:  Boris   |       Owner:  somebody
     Type:  defect  |      Status:  new     
 Priority:  normal  |   Milestone:          
Component:  Other   |     Version:  none    
 Severity:  normal  |    Keywords:          
--------------------+-------------------------------------------------------
 Numpy's ''sum'' function is very slow for 1-D vectors (at least on my pc),
 it is a lot faster to use e.g. the dot product with a ones vector (which
 returns the same result):
 {{{
 t2 = Timer(setup='from numpy import *; c=1.*arange(1,1001)',
 stmt='sum(c)')
 t3 = Timer(setup='from numpy import *; c=1.*arange(1,1001)',
 stmt='c.sum()')
 t4 = Timer(setup='from numpy import *; c=1.*arange(1,1001); d=ones(1000)',
 stmt='dot(c,d)')

 t2.timeit(10000)
 =0.204298445144

 t3.timeit(10000)
 =0.164432998837

 t4.timeit(10000)
 =0.0447709820354
 }}}
 My configuration:
 Windows XP (Pentium D),
 Numpy 1.0.3

-- 
Ticket URL: <http://scipy.org/scipy/numpy/ticket/595>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.