Re: [ublas] += and -= optimized for coordinate_matrix?

From: Gunter Winkler (gunter.winkler_at_[hidden])
Date: 2006-02-15 06:34:30

On Tuesday 14 February 2006 17:30, David M Garza wrote:
> Are the overloaded += and -= operators for coordinate_matrix set up to
> simply append the elements to the end of the matrix arrays?

Looking at matrix_sparse.hpp around line 4232 shows that the += (plus_assign), 
-= (minus_assign) are not optimized in any way. They use the

            matrix_assign<scalar_plus_assign> (*this, ae);

algorithm. 

Personally, I never used these operators. What kind of application do you have 
in mind?

mfg
Gunter