Storage management mechanisms play a significant role in the implementation of modern programming languages. Language facilities for dynamic allocation of data objects and the local environments of procedures require the support of both stack and heap based allocation schemes. One property shared by almost all storage management techniques is their cost decreases as the number of objects decreases even if the amount of memory involved remains constant. The investigator has developed techniques that can reduce the allocation overhead associated with procedure calls by exploiting this property. This is a project in which these techniques will be extended to other domains, including the allocation of memory for procedures in languages requiring retention of activation records and the allocation of dynamically allocated data objects.
