Re: plot problem in 3.4.2 (was: Testing binaries of octave for windows (MinGW32) is updated)

[Liam Groener]

On Jul 25, 2011, at 5:48 AM, Uwe Brauer <address@hidden> wrote:

> 
>> Correction.
> 
>> *******
>> I feel that on windows XP start up time is longer than that in win 7.
>> ********** 
> right it needs some 3 minutes to start up. There are 2
> things I have to mention
> 
>    -  plotting is slow, and the connection to scite works
>       but is even slower!
> 
>    -  there is a problem with a special file which uses
>       isosurface etc. I attach the code below.
> 
> I have compiled version 3.4.2 for ubuntu and I have used
> your precompiled version in XP. For both version I obtain
> the same plot result. BTW the code does not work in 3.2.4!!
> 
> However there are users on this list using version 3.4.2
> officially distributed by suse, they obtain a different
> plot, which is closer to the plot obtained by matlab.
> 
> It is also possible to modify the code such that the plot is
> separated in subplots, these subplots are shown in all
> compiled version.
> 
> I have attached the plot in question already on the list and
> therefore don't want to send it again.
> 
> 
> I am deeply puzzled and would like to ask for help.
> 
> Uwe Brauer 
> 
> Code in question
> 
> graphics_toolkit fltk 
> 
> [x,y,z]=meshgrid(-20:.5:20);
> function Psi=FdOn320(x,y,z)
> r=sqrt(x.^2+y.^2+z.^2);
> Psi=exp(-r./3).*(3*z.^2-r.^2).*2*sqrt(15/(120*pi))/81;
> end
> Psi=FdOn320(x,y,z);
> m=min(Psi(:));
> M=max(Psi(:));
> color=[1 1 0;1 0 1; 0 1 1;0 1 0;1 0 0; 0 0 1; 0 0 0];
> n=2;
> for i=1:n-1
>    isovalue=m+i*(M-m)/n;
>    fv = isosurface(x,y,z,Psi,isovalue);
>    hpatch = patch(fv)
>    isonormals(x,y,z,Psi,hpatch);
>    Alphalevel=0.2+0.8*i/n;
>     set(hpatch,'FaceColor',color(i,:),'EdgeColor','black')
>     daspect([1,1,1])
>     view(3); 
>     axis tight
> end
Just in regards to plot time, I have both 3.4 (the binary prepared by Julien 
Salort) and 3.4.2 installed with MacPorts on my iMac. Running tic;test;toc, 
where test.m is your file, I get 36.3 seconds with octave 3.4 and 1.3 seconds 
with 3.4.2. Actually, in the later case, it takes a moment after the run 
completes for the plot to display; I would guess the elapsed time was more like 
3 seconds. I don't know why the big difference between 3.4 and 3.4.2. I would 
guess some substantial improvement was made in fltk.