"jacob navia" <ja*********@jacob.remcomp.fr> wrote in message news:<bl**********@news-reader5.wanadoo.fr>...
isPow2 = x && !( (x-1) & x );
Algorithm:
If x is a power of two, it doesn't have any bits in common with x-1, since it
consists of a single bit on. Any positive power of two is a single bit, using
binary integer representation.
This means that we test if x-1 and x doesn't share bits with the and operator.
Of course, if x is zero (not a power of two) this doesn't hold, so we add an
explicit test for zero with xx && expression.
Few Mathematicians say any positive integer raised to the power minus
infinity is 0. That is,
n (pow) (-infinity) = 0
IOW, 2 (pow) 3 = 8, 2 (pow) 2 = 4, 2 (pow) 1 = 2, 2 (pow) 0 = 1,
and, 2 (pow) (-infinity) = 0.
Because of this religious reason, I would just write my macro as
(without 0 check)
#define ISPOWOF2( n ) ( ! ( n & (n-1) ) )
But, someone here in CLC told me that the above macro won't work all
the time. I know, it won't work for float; I know the necessity of
additional paranthesis for n. But, I couldn't see any other reasons.
Anybody have any good comments? TIA
---
"Silence is the only right answer for many wrong questions" --
G.K.Moopanaar, Indian Politician
http://guideme.itgo.com/atozofc/ - "A to Z of C" Project
Email: rrjanbiah-at-Y!com