Re: oct function question

[Paul Kienzle]

On Oct 1, 2006, at 2:14 PM, Anand Patil wrote:

> OK: I'm writing lightweight oct functions that get called millions of
> times that involve nonlinear operations on matrix elements. Past
> experience with such situations in C++ indicates that the overhead from
> any function calls or indirection involved in using Matrix's get and 
> set
> functions may add up. Is there any way to get the address of the
> underlying array out of the Matrix class, as with the mex function
> mxGetPr()?

You can get a handle to the underlying data in two ways.

To modify a matrix:

  Matrix data(args(k).matrix_value());
  double *p = data.fortran_vec();

To query a matrix:

  const Matrix data(args(k).matrix_value());
  const double *p = data.data();

John mentioned changing these both to use .data(), with the
const form returning in place and the non-const form performing
a copy, but I don't know if this has been done.

> Also: Is there any way to construct a Matrix so that it shares its
> underlying array with another Matrix? That is, is it possible to pass a
> matrix into an oct function but then work with it under a name other
> than args(i).matrix_value()?

Not sure I understand your question.  The matrix object owns the
data and there is no way to create a new matrix object which points
to existing data.  Two matrix objects can point to the same data,
but the data is copied if either are modified.

- Paul