builtin kron function.

[etienne grossmann]

  Hello,

> The octave-sources mailing list is for posting sources.

  Oops!

> Can you please use the help-octave mailing list for discussion?  Or,
> for the bleeding-edge versions, the octave-maintainers list is
> probably better.                     ^
                                       |
                                       I did not know about it. In
                                       fact it is not advertized on
                                       the ocave sit.
> Thanks,
> jwe

  
  so, as I was writing yesterday,

======================================================================
  Hello,

  I added to 'data.cc' a 'kron2' function that implements the
Kronecker product (assuming real matrices). It came out to be just
barely faster than the 'mykron' function I sent earlier today, I
wonder why.

  The function is built around a double loop, where 'a','b' are the
input and 'c' is the output.

  Matrix c (rc,cc) ;
  for( i=0,ia=-1; i<rc; i++ ) {
    i%rb || ia++ ;
    for( j=0,ja=-1; j<cc; j++ ) {
      j%cb || ja++ ;
      c(i,j) = a(ia,ja) * b( i % rb, j % cb ) ;
    }
  }

  I can use c.elem(i,j) instead of c(i,j), but that does not improve
things. Would there be a way of accessing "more directly" to the
elements? 

  Cheers,

  Etienne


ps : Why does 'make' in "src" cause 'octave' to be rebuilt even if
     nothing has been touched?

pps: Does it sound like a good idea to rewrite the 'reshape' function
     as a builtin?





---------------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.  To ensure
that development continues, see www.che.wisc.edu/octave/giftform.html
Instructions for unsubscribing: www.che.wisc.edu/octave/archive.html
---------------------------------------------------------------------