Bit shifting is much faster than multiplication and especially division. The diffirence is not that huge anymore beacuse multiplication/division is much faster on modern processors than it was before. But still, in a tight loop one can get quite a boost by using bitshifts.

Good compliers suppose to optimize muls/divs into shifts for you but i am not sure that they actually do it.

Bitshifts olny work with integer numbers, not floats.

Here''s an example:
A/2 is same as A>>1
A/4 is same as A>>2
A/8 is same as A>>3
etc
A*2 is same as A<<1
A*16 is same as A<<4
etc

Here''s a tricky one (I am sure complier will not optimize it):
A*320 is same as (A*256) + (A*64) same as (A<<8) + (A<<6)

Cheers

Except for when I''m doing really wacky math sometimes, I notice no difference with my apps. It seems all the mults are being optimized into shifts anyway.

Yeah, i see some speed increases with bit shifting, but all these modern JITs and optimizers and compilers and what not usually change mutliplcation and division by powers of 2 into bit shift ops.

this is how it works:

2 * 4 = 8 (also) 2 * (2^2) = 8
0010 * 0100 = 1000
(you can''t really visualize the math
with binary, but that''s what 2,4,& 8
are in binary)

and
2 << 2 = 8
0010 (move everything right by 2) = 1000

Bit shifting is also usefull for dealing with numbers that don''t represent numbers.
(i had this really long explanation but it got too confusing)
Look up bit masking online, or in a book