Next: Building and Dismantling Terms, Previous: The Determinacy Checker, Up: Writing Efficient Programs [Contents][Index]
Another important efficiency feature of SICStus Prolog is last call optimization. This is a space optimization technique, which applies when a predicate is determinate at the point where it is about to call the last goal in the body of a clause. For example,
% for(Int, Lower, Upper) % Lower and Upper should be integers such that Lower =< Upper. % Int should be uninstantiated; it will be bound successively on % backtracking to Lower, Lower+1, … Upper. for(Int, Int, _Upper). for(Int, Lower, Upper) :- Lower < Upper, Next is Lower + 1, for(Int, Next, Upper).
This predicate is determinate at the point where the
recursive call is about to be made, since this is the last clause
and the preceding goals (<)/2
and is/2
) are
determinate. Thus last call optimization can be applied;
effectively, the stack space being used for the current predicate
call is reclaimed before the recursive call is made. This means that
this predicate uses only a constant amount of space, no matter how
deep the recursion.
• Accumulating Parameters | Accumulating Parameters | |
• Accumulating Lists | Accumulating Lists |