Marking algorithms

Abstract

Algorithms for traversing and marking the nodes of a directed graph have applications in many fields, for instance search inethods in artificial intelligence and garbage collection schemes. In this paper, a general nonrecursive algorithm for the purpose is formulated and proved, and some if its properties are investigated. A second general nonrecursive algorithm is also discussed. Then two implementations of the general algorithms with valuable properties are described. Finally a recursive version is given.

A fast compactifying garbage collector

“…Wegbreit [9] assumes pointers to locations within blocks, which entails marking of singular words. The compaetifying garbage collector in the LAX/DPL5 system mentioned in [6] is inferior to the present method: it uses one more word per string and one more phase. The compaetifying garbage collector in the LAX/DPL5 system mentioned in [6] is inferior to the present method: it uses one more word per string and one more phase.…”

“…Let The general case can be translated to the tree case by using the notion of the mark tree [6], defined by following the pointers in depth-first order while discarding pointers to blocks already reached.…”

Global garbage collection for distributed heap storage systems

“…In this case, the local roots are RQ, SQ, and all ODT entries that have their Ext fields set. We use a sequential marking algorithm (17) for marking local objects and use the same local sweep and compaction algorithm described in Section 13 and in Appendix D. …”