Leaf subroutine
A leaf subroutine is a subroutine which cannot in turn call another subroutine. Some compilers can apply special program optimizations to leaf subroutines to make them more efficient, such as the use of link registers to avoid having to push the return address on the stack, or not allocating a register window on CPU architectures descended from Berkeley RISC.[1]
The term "leaf" refers to their position as leaf nodes in the call graph of the program.
Usually, most non-leaf subroutines call more than one other subroutine. When this is the case, the majority of subroutine calls in the call graph are calls to leaf subroutines, because a binary tree has more leaf nodes than non-leaf nodes (assuming that all non-leaf nodes have two children). Consequently, the efficiency of calls to leaf subroutines often has a significant effect on the efficiency of the whole program. [citation needed]
References
[edit]