[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Implementation of convn
From: |
Søren Hauberg |
Subject: |
Re: Implementation of convn |
Date: |
Mon, 24 Mar 2008 22:22:07 +0100 |
man, 24 03 2008 kl. 16:27 -0400, skrev John W. Eaton:
> Will you please update according to the following comments and submit
> this as an hg changeset?
Sure. In which directories do you want the files to be placed?
> | // For each neighbour
> | SumType sum = 0;
> | for (octave_idx_type n = 0; n < ndims; n++)
> | B_idx(n) = 0;
> | for (octave_idx_type j = 0; j < B_numel; j++)
> | {
> | for (octave_idx_type n = 0; n < ndims; n++)
> | A_idx(n) = out_idx(n) + (B_size(n)-1-B_idx(n));
> | sum += A(A_idx)*B(B_idx);
> | B.increment_index(B_idx, B_size);
> | }
> |
> | // Compute filter result
> | out(out_idx) = sum;
> |
> | // Prepare for next iteration
> | out.increment_index(out_idx, out_size);
> |
> | OCTAVE_QUIT;
> | }
> |
> | return octave_value(out);
> | }
>
> Is there a way to do this without increment_index? That function will
> make this code slow.
Perhaps. I couldn't think of a way though. Perhaps somebody else has a
suggestion?
> Note also that doc strings for internal functions
> should just say (for example)
>
> "-*- texinfo -*-\n\
> @deftypefn {Loadable Function} {} __convn__ (@var{a}, @var{b})\n\
> Undocumented internal function"
>
> Currently we have a mixture of styles for these, so maybe they should
> all be fixed to be consistent.
Don't we have "private" directories now? I would think all __functions__
should be moved to "private" directories. If we do this, then it would
probably be nice for developers if __functions__ are documented as well.
Søren
- Implementation of convn, Søren Hauberg, 2008/03/23
- Re: Implementation of convn, Søren Hauberg, 2008/03/23
- Re: Implementation of convn, John W. Eaton, 2008/03/24
- Re: Implementation of convn,
Søren Hauberg <=
- Re: Implementation of convn, John W. Eaton, 2008/03/24
- Re: Implementation of convn, Søren Hauberg, 2008/03/24
- Re: Implementation of convn, John W. Eaton, 2008/03/24
- Re: Implementation of convn, Søren Hauberg, 2008/03/25
- Re: Implementation of convn, John W. Eaton, 2008/03/25
- Re: Implementation of convn, Søren Hauberg, 2008/03/26
- Re: Implementation of convn, John W. Eaton, 2008/03/26
- Re: Implementation of convn, Søren Hauberg, 2008/03/27
- Re: Implementation of convn, David Bateman, 2008/03/27
- Re: Implementation of convn, John W. Eaton, 2008/03/27