cos/sin acting strange?
Author
This is probably something really obvious I''m overlooking, but I don''t see it. Is there some attribute of the c/c++ standard library trig functions?
Heres the problem.
Why does this: cosf(acosf(0))
Give me this: -4.37114e-08
cosf gives me that value when I feed in a right angle. Am I stupid or is cos(PI/2) not 0? I''ve tried this in linux and VCPP.
Also if I do this:
float HALFPI = M_PI / 2;
float value = sinf( HALFPI - acosf(0) );
value will be 0.
But if I do this:
float value - sinf( (M_PI/2) - acosf(0) );
value will again be -4.37114e-08
Does this make any sense to anyone?
It's caused by floating point inaccuracy. For more information, read up on the IEEE 754 standard floating-point format.
Don't worry though; for all intents and purposes, that number IS zero, except for the fact that its binary representation is not literally zero, so you'd have to check its absolute value against a suitable epsilon to see if it was small enough to be considered zero.
Later,
ZE.
//email me.//zealouselixir software.//msdn.//n00biez.//
miscellaneous links
[edited by - zealouselixir on July 8, 2003 12:21:24 AM]
Don't worry though; for all intents and purposes, that number IS zero, except for the fact that its binary representation is not literally zero, so you'd have to check its absolute value against a suitable epsilon to see if it was small enough to be considered zero.
Later,
ZE.
//email me.//zealouselixir software.//msdn.//n00biez.//
miscellaneous links
[edited by - zealouselixir on July 8, 2003 12:21:24 AM]
Why use float? Why not use double and use the sin and cos functions?
And yes, -4.37114e-08 is a very small number, close enough to zero.
And yes, -4.37114e-08 is a very small number, close enough to zero.
<a href="http://docs.sun.com/source/806-3568/ncg_goldberg.html">What Every Computer Scientist Should Know About Floating-Point Arithmetic</a>
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement