[SciPy-user] numpy install errors

[Nils Wagner]

On Mon, 18 Feb 2008 11:48:33 -0600
  "Robert Kern" <robert.kern at gmail.com> wrote:
> On Feb 18, 2008 11:42 AM, Alastair Basden 
><a.g.basden at durham.ac.uk> wrote:
>> Hi,
>> installation now okay - was something to do with 
>>environment variables,
>> though not sure which ones...
>>
>> So the question now is why does the following happen:
>>
>> >>> import scipy.special
>> >>> scipy.special.kv(6./5,1)
>> 0.70066931017889988
>> >>> scipy.special.kv(6./5,1)
>> 0.70066931017889988
>> >>> scipy.special.kv(6./5,1)
>> 0.70066931017889988
>> >>> scipy.special.kv(6./5,[1,1])
>> array([ 0.70066931,  0.70066931])
>> >>> scipy.special.kv(6./5,1)
>> 6.1853203003937157e-282
>> >>> scipy.special.kv(6./5,1)
>> 6.1853203003937157e-282
>>
>> ie it seems to go wrong after the first ufunc is 
>>called... but is okay on
>> single values...
> 
> There's a bug somewhere. A similar problem has been 
>reported recently.
> 
> -- 
> Robert Kern
> 
  
I cannot reproduce the problem with recent svn versions.

  Python 2.4.1 (#1, May 25 2007, 18:41:31)
[GCC 4.0.2 20050901 (prerelease) (SUSE Linux)] on linux2
Type "help", "copyright", "credits" or "license" for more 
information.
>>> import numpy,scipy.special
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(11./6,numpy.ones((10,),numpy.float64))
array([ 1.32620202,  1.32620202,  1.32620202,  1.32620202, 
 1.32620202,
         1.32620202,  1.32620202,  1.32620202, 
 1.32620202,  1.32620202])
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,1)
0.70107989955789207
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.special.kv(6./5,[1,1])
array([ 0.7010799,  0.7010799])
>>> scipy.__version__
'0.7.0.dev3946'
>>> numpy.__version__
'1.0.5.dev4811'
>>> scipy.special.kv(6./5,1)
0.70107989955789207


Nils