The A64 and FX line of "brainiacs" could process upwards of three instruction per clock cycle. While the P4 "speed demon" had something around a 1 IPC rate. Add to that the AMD chips has a much more efficient architecture and an on-die memory controller, They are undoubtedly the best of the single core chips.
Intel saw how efficient this low clock rate, high IPC rate was, and adopted it into their next series of chips, the core 2 duo's. They improved their micro-architecture around this point, and created a chip AMD has trouble with beating to this day. They're even going to try the on-die memory controller approach with the Nehalem series. Bunch of copy-cats..