Interesting fact:
I love the 8051, worked with for many years, and still working, but, when you comparing the above 8051 clock cycles to the AVR, this one can rotate one bit in its internal 32 accumulators tied up to the ALU, in a SINGLE clock cycle, that means the AVR does it at least 100 times faster than the plain 8051.
One of the AVR tricks is that it fetches memory PROGRAM in 16 bits wide, so instruction (2 bytes) are fetched at once, while in the 8051 it takes two machine cycles (24 clock cycles).