Adapting Barrel Shifter at Compilation Level for Efficient Implementation of Multiplications

Multiplication is one of the most studied implementations in computing. Some architectures implement it as a single operation, while some others implement it as a combination of other operations. A typical implementation is the repeated addition method, where the operands are repeatedly added to get the result. Here, we try to modify this implementation by using the barrel shifter. The barrel shifter can shift block of bits either to the left or the right, and append zeroes at the end. We can exploit the distributive nature of multiplication over addition to convert the multiplication operation to a series of shift operations, reducing time to a good degree. For large operands, it can save even 85% of the time.