---

Introduction: leaf and non-leaf subroutines

---

• Leaf subroutine
  • Leaf subroutine (leaf function):

    Leaf subroutine/function = a subroutine/function that does not contain a subroutine/function call

    Example:

    public class myClass { public int square(int x) // square( ) is a leaf function { return x * x; } ... }

  ---

  ---

  ---

  ---

• Non-leaf subroutine
  • Non-leaf subroutine (leaf function):

    Non-leaf subroutine/function = a subroutine/function that contains one or more subroutine/function calls

    Example:

    public class myClass { public int square(int x) // square( ) is a leaf function { return x * x; } public int sumOfSquare(int x, int y) // sumOfSquare( ) is a non-leaf function { return square(x) + square(y); } ... }

  ---

  ---

  ---

  ---

• Why we make this distinction ?
  • It's common to have many more leaf subroutines in a high level language program:

    Programmers that practice modular programming tends to write small subroutines to do a (small but) specific task

  • Some modern processors (like the ARM processor) provide a very efficient subroutine call instruction in assembler that only a leaf subroutine can use...

    The programmer must manage the return address in order to implement non-leaf subroutines !!!

    Older processors (like the INTEL x86 and Core Dual) has a subroutine call instruction for non-leaf subroutines that a leaf subroutine can also use.

    So:

    The assembler programmer is more inconveniencedd by the ARM processer. That's OK. It's a small inconvenience (to gain (a little) performance)