[NumPy-Tickets] [NumPy] #1666: Reducing the memory usage of MaskedArrays after DomainedBinaryOperation

NumPy Trac numpy-tickets@scipy....
Thu Nov 11 12:33:57 CST 2010


#1666: Reducing the memory usage of MaskedArrays after DomainedBinaryOperation
-------------------------------+--------------------------------------------
 Reporter:  lee                |       Owner:  pierregm   
     Type:  enhancement        |      Status:  new        
 Priority:  lowest             |   Milestone:  Unscheduled
Component:  numpy.ma           |     Version:  none       
 Keywords:  mask memory usage  |  
-------------------------------+--------------------------------------------

Comment(by lee):

 Thanks for the reply. I think r6335 does most of what I wanted. I had a
 case where I take the modulus of a masked array (with {{{nomask}}}), and
 get a fully masked array in return, even though a cursory check of the
 floating point number would tell me that there is no way to get an
 infinite number (divide-by-zero) using this floating point number.

 Checking the floating point number ({{{b}}}) prior to the modulus/division
 may be a better idea  that would speed up the operation when b=0 (no
 division necessary, just set to Inf and mask it), and when b!=0 (division
 necessary, but no additional masking).

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1666#comment:3>
NumPy <http://projects.scipy.org/numpy>
My example project


More information about the NumPy-Tickets mailing list