[Numpy-discussion] floating point arithmetic issue

Guillaume Chérel guillaume.c.cherel@gmail....
Fri Jul 30 07:21:23 CDT 2010


  Hi,

Thanks for all your answers and the references (and yes, I have to admit 
that I've been a bit lazy with Goldberg's article, though it looks very 
thorough).

But as numpy is designed for scientific computing, is there no 
implementation of an "exact type" 
(http://floating-point-gui.de/formats/exact/) to avoid floating point 
issues?

As for the details about my problem, I'm trying to compute the total 
surface of overlapping disks. I approximate the surface with a grid and 
count how many points of the grid fall into at least one disk. I use the 
mod operator in a calculation to find the coordinates of the left-most 
point of the grid that falls into a given disk, and I do that for all 
disks. I need to use floating point numbers with this modulo operator 
because I want the resolution of the grid is arbitrary.

I suppose my explanation might not be very straightforward, so if you 
know of another good way to compute the surface of overlapping disks, 
I'd be glad to know.

Thanks,
Guillaume

Le 30/07/2010 13:21, Pauli Virtanen a écrit :
> Fri, 30 Jul 2010 19:52:37 +0900, David wrote:
> [clip]
>> Indeed, it is not, and that's expected. There are various pitfalls using
>> floating point. Rational and explanations:
>>
>> http://docs.sun.com/source/806-3568/ncg_goldberg.html
> In case of tl;dr, see also http://floating-point-gui.de/
>


More information about the NumPy-Discussion mailing list