[SciPy-user] conj and element-by-element multiplication

[Scott Ransom]

Hi David,

You can prevent the upcasts using the .savespace() method.  For
example:

In [8]:a = arange(10.0, typecode='F')

In [9]:a**2
Out[9]: array([  0.+0.j,   1.+0.j,   4.+0.j,   9.+0.j,  16.+0.j,
25.+0.j, 36.+0.j,  49.+0.j,  64.+0.j,  81.+0.j])

In [10]:a.savespace()

In [11]:a**2
Out[11]: array([  0.+0.j,   1.+0.j,   4.+0.j,   9.+0.j,
16.+0.j,  25.+0.j, 36.+0.j,  49.+0.j,  64.+0.j,  81.+0.j],'F')

This is _very_ useful when dealing with very large arrays...

Scott

> Curious. One more question: are these Complex32 matrices? I.e., are  
> they matrices using C floats (single precision), as opposed to C  
> doubles (double precision)? The reason is that if x is a vector of  
> floats, x**2 upcasts x to doubles first (b/c 2 is an int -- ints are  
> cast to doubles), so that kills the speed, and I can get timings like  
> this. This isn't a good behaviour, and I'm going to see what I can do  
> to fix this for the next version of Numeric.
> 

-- 
Scott M. Ransom            Address:  NRAO
Phone:  (434) 296-0320               520 Edgemont Rd.
email:  sransom at nrao.edu             Charlottesville, VA 22903 USA
GPG Fingerprint: 06A9 9553 78BE 16DB 407B  FFCA 9BFA B6FF FFD3 2989