A Multi-Processor Shared Memory Architecture

As an alternative to mark-scan algorithms Lins presents a shared memory architecture for parallel cyclic reference counting, based on the algorithm for cyclic reference counting with local and lazy mark-scan. In this paper we generalise this architecture in such a way that multiple mutators and collectors share the same workspace. This generalisation is simple and keeps the properties of the one-mutator-one-collector architecture. (to appear in Microprocessing & Microprogramming - Aug/92)