The Brian Kernighan approach isn't mentioned: it involves clearing the lowest set bit automatically, so the loop time is related to the number of set bits(n), not the number of bits (8).

http://www.keil.com/support/docs/194.htm


and for parity here is how to get it for a byte in C (using Keil):
http://www.keil.com/support/docs/1620.htm

to do a 16 bit word you add the parity bits of the 2 individual bytes.