1986/07/01 by Robert Bernstein
Computer Science · #Parallel Computing and Optimization Techniques #Embedded Systems Design Techniques #Numerical Methods and Algorithms
paper · doi:10.1002/spe.4380160704
Abstract Methods are given for finding a sequence of ‘add’, ‘subtract’ and ‘shift’ instructions to multiply the contents of a register by an integer constant. Each method generalizes the previous one and requires only a few intermediate or scratch registers. A variation of the last method is used in the PL.8 compiler and uses an unnoticeable amount of the overall compile time. Some statistics roughly indicating the effectiveness of the methods are presented.