Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Distributed resource management: garbage collection

Thesis/Dissertation ·
OSTI ID:7183929

In recent years, there has been a great interest in designing high-performance distributed symbolic-processing computers. These architectures have special needs for resource management and dynamic reclamation of unused memory cells and objects. The memory management or garbage-collection aspects of these architectures are studied. Also introduced is a synchronous distributed algorithm for garbage collection. A special data structure is defined to handle the distributed nature of the problem. The author formally expresses the algorithm and shows the results of a synchronous garbage-collection simulation and its effect on the interconnection-network message to traffic. He presents an asynchronous distributed garbage collection to handle the resource management for a system that does not require a global synchronization mechanism. The distributed data structure is modified to include the asynchronous aspects of the algorithm. This method is extended to a multiple-mutator scheme, and the problem of having several processors share portion of a cyclical graph is discussed. Two models for the analytical study of the garbage-collection algorithms discussed are provided.

Research Organization:
Texas Univ., Austin (USA)
OSTI ID:
7183929
Country of Publication:
United States
Language:
English