Re: Problem with plotting of functions for large arguments

[Przemek Klosowski]

On 10/30/2015 05:09 AM, kschnikow wrote:

Dear all,

I have a problem using the plot() functions for e.g. plotting the sine
function for large arguments  > 1e6

For example, when I define the intervall x=10000000*pi:4*pi/100/10000004*pi;
and I try to plot(x,sin(x)) I get rubbish.

The interval above must be a typo: you probably meant to say

x=10000000*pi:4*pi/100:10000004*pi

Now, it's true that the resulting plot(x,sin(x)) is wrong. The values are actually correct, as shown by plotting them as points by plot(x,sin(x),'.').

This is because the default graphics toolkit, fltk, has issues with plotting large magnitude numbers because uses single precision floating point internally. Most Octave installs will have two graphics toolkits available: fltk and gnuplot, and gnuplot does not have this problem, so switch to it and the plot will come out fine:

graphics_toolkit ('gnuplot')
 plot(x,sin(x))