modulous

Modulo is an integer only operation. It gives you the remainder of an integer division. There is no remainder for floating point divisions, only an error term. That''s why the % operator will not work on floating point types.

thanks

isn''t there a "floating point" modulus?

float fModulous = (int) fA % (int) fB;

quote:

---

isn't there a "floating point" modulus?

---

Nope. The modulus operation makes no sense from a mathematical point of view, when applied to a floating point number. The modulus is defined as the remainder of a division. In a floating point division, a remainder does not exist (or is always zero). So mathematically expressed: float_1 % float_2 = 0. The result would always be zero.

The fact that: (float_1 / float_2) * float_2 != float_1 is due to floating point inaccuracies, but not to a modulus (as it would be if both value were integers).


[edited by - Yann L on December 28, 2002 8:33:02 AM]

Well, I''ve used this in my code, and it works just fine

fmod is not really the same as integer modulus. fmod(x,y) returns x-(a*y), where a is the largest integer, so that the result has a smaller magnitude, though the same sign as x. In some cases it can simulate an % operator, in some not.

The mod operation for floating point numbers is defined as:

x mod y = x - y * floor(x/y) ,

where floor() returns the greatest integer less than or equal to the given number.

So, if x = 1.0 and y = 0.33:

1.0 mod 0.33 = 1.0 - 0.33 * floor(1.0/0.33)
= 1.0 - 0.33 * 3.0
= 1.0 - 0.99
= 0.01

It's worth noting that x = q*y + (x mod y) , where q is an integer value (the quotient) that is equal to floor(x/y).

So, while it's true that C's % operator doesn't operate on floating point numbers, it's nevertheless possible to define the mod operation for all real numbers, as above.

(Source: The Art of Computer Programming, Vol. 1 by Donald Knuth)

[Edit: I originally said that floor() drops the number's fractional part, but that's not true when the number is negative. Also, I forgot "or equal" above. Sorry.]

[edited by - chronos on December 28, 2002 4:24:30 PM]

Another thing about C''s % operator is that, as far as I know, its behavior is undefined when either operand is negative, so C''s % operator may or may not correspond to the mathematical modulo operator.

Mathematically, the result of -6 mod 5 is 4. In C, the result of -6 % 5 is either -1 or 4, depending on the implementation.

Pretty confusing, isn''t it?