Dynamic memory allocation has been a fundamental part of most computer systems since roughly 1960, and memory allocation is widely considered to be either a solved problem or an insoluble one. In this survey, we describe a variety of memory allocator designs and point out issues relevant to their design and evaluation. We then chronologically survey most of the literature on allocators between 1961 and 1995. (Scores of papers are discussed, in varying detail, and over 150 references are given.) We argue that allocator designs have been unduly restricted by an emphasis on mechanism, rather than policy, while the latter is more important; higher-level strategic issues are still more important, but have not been given much attention. Most theoretical analyses and empirical allocator evaluations to date have relied on very strong assumptions of randomness and independence, but real program behavior exhibits important regularities that must be exploited if allocators are to perform well in practice.

Smalltalk, Lisp, or actor languages typically violate some of the basic lo- Appears in Proc. 1992 ACM Conf. on Lisp and Functional Programming, San Francisco, California, June 2224, 1992, pp. 32--42. This and other papers on garbage collection, memory hierarchies, and persistence are available via anonymous internet ftp from cs.utexas.edu, in the directory pub/garbage. cality assumptions of modern memory hierarchies, leading to poor performance of normal caching strategies. The main problem is cyclic reuse of memory at a time scale too long to be captured by any caching policy. These systems therefore tend toward being bandwidthlimited, i.e., dependent on the speed of transfers between levels of the memory hierarchy. 2.1 Allocation is the problem, not garbage collection per se The problem of locality in garbage collected systems does not depend primarily on the locality of the garbage collectio

We present a new abstract machine for graph reduction called TIGRE. Benchmark results show that TIGRE's ex- ecution speed compares quite favorably with previous combinator -graph reduction techniques on similar hardware. Furthermore, the mapping of TIGRE onto conventional hardware is simple and efficient. Mainframe implementations of TIGRE provide performance levels exceeding those previously available on custom graph reduction hardware.