32 bits processor optimized code

[Robertus]

I have a test in front of me. It is about parsing an unsigned char* string, BIT by BIT and I must tell at each call of a certain method, wich is the next "changing element" in that bit string; a "changing element" is in fact a bit with its value different from the value of the previous bit:

00010001 --> here the 4-th and 5-th and 8-th bits are "changing elements"

One of the requests is that I should implement the algorithm in a manner that would generate 32-bit processor optimized code. WHAT DO THEY MEAN WITH THAT???? HOW CAN I DO IT???

Please, help me!