A piece of code that is obeyed ‘out of line’, i.e. control is transferred to the subroutine, and on its completion control reverts to the instruction following the call. (The instruction code of the CPU usually provides subroutine jump and return instructions to facilitate this operation.) A subroutine saves space since it occurs only once in the program, though it may be called from many different places in the program. It also facilitates the construction of large programs since subroutines can be formed into libraries for general use. (The same concept appears in high-level languages as the procedure.)

In the early days of programming, what is now called a subroutine was known as a closed subroutine. This was in contrast with the open subroutine, which was a piece of code that appeared in several places in a program, and was substituted ‘in line’ by the assembler for each call appearing in the program. The open subroutine was just a convenient shorthand for the programmer: the same facility is now known as a macro.