fast hist() / histogram() implementation in C++

[Petr Mikulik]

It seems that the M-language does not provide an efficient way to implement
histogram function. It is order(s) of magnitude worse than  ndat*log2(nbins)
efficiency of a loop-friendly language, like C/C++. As an example, try the
following histograming code for a 2048x2048 image:
        n=2048;
        nbins=100;
        a=1./hilb(n);
        imagesc(a);
        tic; [y,x]=hist(a(:),nbins); toc


I'm using my own function called "histogram" with this syntax:
  [ [x, ] h ] = histogram ( image [, bins [, zfrom, zto] [, do_log ]] )

I've put its M and well as C++ (=>.oct) code to
        http://www.sci.muni.cz/~mikulik/tmp/pmhist4octave.zip

It would be great if someone uses its code (...or his own code) for a C++
(thus => .oct routine) implementation of
        histc(data, edges)
function for calculating histogram based on given vector of bin edges. Then,
hist() function should call this histc.oct instead of its current inefficient
M-language implementation.

Looking forward to histc.cc.

Petr Mikulik



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------