 
Summary: A Time Complexity Bound for Adaptive Mutual
Exclusion ?
(Extended Abstract)
YongJik Kim and James H. Anderson
Department of Computer Science
University of North Carolina at Chapel Hill
Abstract. We consider the time complexity of adaptive mutual exclu
sion algorithms, where \time" is measured by counting the number of
remote memory references required per criticalsection access. We estab
lish a lower bound that precludes a deterministic algorithm with O(log k)
time complexity (in fact, any deterministic o(k) algorithm), where k is
\point contention." In contrast, we show that expected O(log k) time is
possible using randomization.
1 Introduction
In this paper, we consider the time complexity of adaptive mutual exclusion
algorithms. A mutual exclusion algorithm is adaptive if its time complexity
is a function of the number of contending processes [3, 6, 8, 10, 11]. Under the
time complexity measure considered in this paper, only remote memory refer
ences that cause a traversal of the global processortomemory interconnect are
counted. Specically, we count the number of such references generated by one
