Re: [slightly OT]: Octave featured in Free Software magazine

[David Bateman]

Alexander Barth wrote:
> Hi Jordi,
> Thank you for pointing this benchmark out.
> 
> Octave's speed can be improved quite significantly if lines like:
> 
>     qgtfive= find(qRe > 5.);
>     zRe(qgtfive)=5.;
> 
> are replaced by:
> 
>     zRe(qRe > 5.)=5.;
> 
> The code is also much clearer this way.
> On my machine (AMD Opteron 244):
> 
> octave:1> tic; fractal_benchmark; toc
> Elapsed time is 24.935944 seconds.
> octave:2> tic; fractal_benchmark2; toc
> Elapsed time is 14.843300 seconds.

I got much the same results for this

> 
> In addition to this, the complex type can be used instead of real and 
> imaginary part separately.

though this doesn't help as the bizarre limit separately on the real and
imaginary parts of z to be in [-5,5] mean that the you have to go back
to real and imaginary parts to apply the limits to z..

D.