[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: matlab code: isovalues and isosurface

From: Liam Groener
Subject: Re: matlab code: isovalues and isosurface
Date: Sat, 16 Jul 2011 03:46:13 -0700

On Jul 16, 2011, at 2:43 AM, Martin Helm wrote:

Am Freitag, 15. Juli 2011, 23:53:44 schrieb Liam Groener:
(2) What I find very confusing is that the failure occurs almost instantly
when the for loop is run from 1 to n, but the code runs for more than a
minute (without failue) if the loop is run from 1 to n-1. It seems like it
is actually running it from n to 1 in the former case?

I guess you are using gnuplot as backend? You did not mention it. The reason
that the error comes almost instantly is that the loop is very fast, it is
finished long before gnuplot starts to show the first plot. So it has nothing to
do with the loop running in the wrong order as you suspected but seems to be  
simply a side effect of the asynchronous behaviour of the plots running in a
separate process while the calculations are all finished.

Yes, you are right. I was fooled by the facts that all the plotting commands are inside the loop and that most of the cpu time was being used by the octave process rather than by gnuplot. I guess that during the loop, the gnuplot commands are placed in a que and octave continues feeding them to gnuplot over the length of the run?

I have been running with both the gnuplot and fltk backends. I swear I was getting similar plots from both earlier in the day (actually yesterday now). However, recently, I have not been able to make fltk work. It just plots the axes and finishes with no actual plot or error message. The m-file I'm currently using reads:

function Psi=FdOn320(x,y,z)
 color=[1 1 0;1 0 1; 0 1 1;0 1 0;1 0 0; 0 0 1; 0 0 0];
 for i=1:n-1
     fv = isosurface(x,y,z,Psi,isovalue);
     hpatch = patch(fv);
      axis tight
 %     camlight left; 
 s = 'hi'

Earlier, I was trying small variations in this. Note that I uncommented out the daspect command since octave 3.4.2 supports it. I just added the last line, my crude way of verifying that the loop finishes fast as you said.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]