Is addition as fast as bit wise operations in modern processors,is there any other optimization after full adder


#1

5.Why is addition as fast as bit-wise operations in modern processors?


#2

@Pritika_Singh Addition in our modern day processors are still based on a basic adder circuit, but they are way more complex and efficient, with usage of ripple carry techniques, or carry look ahead techniques, logic adders, carry save adders. The development of techniques to add more efficiently, is limited by the current architecture being used. Now if today an architecture uses x no. of transistors to perform addition, it produces an output y every z cycles, then maybe in the future we have another architecture, which may use x^8 transistors which will produce an output y every z/8. While full adders are of various types, the simultaneous addition of multiple numbers are limited by the number of transistors present in a processor. For now we also have MAC (multiply-accumulate units) or FMAC(fused multiply accumulate units) which contain a multiplier implemented in combinatorial logic followed by adders and an accumulator register to store the results. the output is produced in 6 cycles whereas other full adder units may do the same operation for a 32 bit data in 32 cycles or more.