Virtual memory is the simulation of a storage space so large that programmers do not need to reprogram or recompile their works when the capacity of a local memory or the configuration of a network changes. The name, borrowed from optics, recalls the virtual images formed by mirrors and lenses--images that are not there but behave as if they are. The designers of the Atlas Computer at the University of Manchester invented virtual memory in the 1950s to eliminate a looming programming problem: planning and scheduling data transfers between main and secondary memory and recompiling programs for each change of size of main memory. Virtual memory is even more useful in the computers of the 1990s, which have more things to hide-on-chip caches, separate RAM chips, local disk storage, network file servers (q.v.), large numbers of separately compiled program modules, other computers on the local bus or local network, or the Internet. The story of virtual memory from then to now is a story about machines helping programmers solve problems in storage allocation, protection of information, sharing and reuse of objects, and linking of program components. Virtual memory, common in all computers and operating systems from the smallest microprocessor to the largest supercomputer, is now invading the Internet.

Virtual memory

This paper reviews virtual storage and virtual machine concepts, consolidating and updating earlier discussions. The manner in which actual virtual storage and machine systems have been implemented, and certain problems of current implementations, are described. To better illustrate the material, the virtual machine system CP-67 for the IBM System/360 Model 67 is considered at some length. An annotated bibliography is included.

Virtual storage and virtual machine concepts

For many years, there has been interest in "program locality" as a phenomenon to be considered in storage allocation. This notion arises from the empirical observation that it is possible to run a program efficiently with only some fraction of its total instruction and data code in main storage at any given time. That virtual memory systems can be made to run at all demonstrates that program locality can be used to advantage; and though it is certainly possible to write a program which violates the principles of locality, it seems one must go out of one's way to do so.

Experiments with program locality

*Partial support for the author's research in memory hierarchies has been provided by the National Science Foundation under grants MCS75,0676~ and MCS77-28429 and the Department of Energy under contracts W-7405-ENG'4~ (to LBL) and EY-7b-C-03-0515 (to SLAC). **Computer Science Division, EECS Department, University of California, Berkeley, 94720. The author is also on the staff of the Lawrence Berkeley Laboratory and is a visitor at the Stanford Linear Accelerator Center.

Bibliography on paging and related topics

The Use and Performance of Memory Hierarchies: A Survey

This short note focuses attention upon techniques for dynamic memory allocation in multiprogrammed systems which employ the addressing mechanisms of paging and segmentation. Here, event indicators and mathematical tools are presented which supply characterizations of the paging and segmentation addressing processes. It is shown that these statistical characterizations form data bases which can be used to derive Bayesian storage allocation algorithms conditionally based upon usage, demand, and processing history. It is argued that these characterizations, and algorithms similar to those constructed here, provide a flexible basis for efficient memory management in multiprogrammed, and by extension, time-shared environments. Although emphasis is directed to managing primary (main) memory residence, the techniques could be extended to govern memory management for a hierarchy of storage devices.

Someshwar

Papers

On the Design of Bayesian Storage Allocation Algorithms for Paging and Segmentation