Prev: Passing structure of pointers as an argument
Next: Is compiling CUDA-based S-functions, the using Real-Time Workshop Embedded Coder possible?
From: matlab user on 24 Mar 2010 20:46
Hello, I have a simple question:

Where does the error -8.8818e-016 come from? thank you in advance.

>> aa=[-0.6033 0.4265 0.9790 -0.1211];
>> bb=[10 10 10 10 ];
>> cc=aa*bb'

cc =
6.8110

>> cc - 6.811

ans =
-8.8818e-016
From: Nicholas Kinar on 24 Mar 2010 21:01
On 24/03/2010 6:46 PM, matlab user wrote:
> Hello, I have a simple question:
>
> Where does the error -8.8818e-016 come from? thank you in advance.
>
>>> aa=[-0.6033 0.4265 0.9790 -0.1211];
>>> bb=[10 10 10 10 ];
>>> cc=aa*bb'
>
> cc =
> 6.8110
>
>>> cc - 6.811
>
> ans =
> -8.8818e-016

Perhaps this error can be explained most succinctly by the inability of
a computer to perform exact mathematics on floating point numbers.
Here's some more information.

http://en.wikipedia.org/wiki/Floating_point_numbers

That's also the reason why you need to be careful when comparing
floating point numbers:

http://www.cygnus-software.com/papers/comparingfloats/comparingfloats.htm

Nicholas
From: Matt Fig on 24 Mar 2010 21:23
http://matlabwiki.mathworks.com/MATLAB_FAQ#Why_is_0.3-0.2-0.1_not_equal_to_zero_.28or_similar.29.3F
 | 
Pages: 1
Prev: Passing structure of pointers as an argument
Next: Is compiling CUDA-based S-functions, the using Real-Time Workshop Embedded Coder possible?