The ARM Cortex-M4 has a very fast multiply instruction (MUL) that executes in a single clock cycle.

However, the MUL instruction requires that all operands be in registers, so another instruction is needed to load the constant into a register, and thus a minimum of two registers and two clock cycles are required.

Sometimes multiplication by a constant can be accomplished with a single instruction using only one register and one clock cycle. For example, shifting a register left by N bits will multiply its content by 2^N. Additional constant multipliers are possible by incorporating the shift within a single ADD, SUB, or RSB instruction.

Although two-instruction replacement sequences are never faster than using a MUL instruction, they can often eliminate the need for a second register.

(Click anywhere to remove this message.)