You can loop over the range and then do result ^= i ^ A[i]. If adapters are slow you don't need them here.

Having only one loop gives you a better memory access pattern, because it's 2 XOR operations in between each memory access. Two loops is the same number of instructions, but it spends one loop ignoring memory and then another loop doing rapid-fire memory accesses. In python there's enough overhead that it's unlikely to matter. But in a faster language running on a busy machine that could make a real difference.