A program's working set is the collection of segments (or pages) recently referenced. This concept has led to efficient methods for measuring a program's intrinsic memory demand; it has assisted in undetstanding and in modeling program behavior; and it has been used as the basis of optimal multiprogrammed memory management. The total cost of a working set dispatcher is no larger than the total cost of other common dispatchers. This paper outlines the argument why it is unlikely that anyone will find a cheaper nonlookahead memory policy that delivers significantly better performance.

/pdf/working-sets-past-and-present-2e92gwwuj5.pdf

Working Sets Past and Present

*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 working-set concept is extended for programs that reference segments of different sizes The generalized working-set policy (GWS) keeps as its resident set those segments whose retention costs do not exceed their retrieval costs The GWS is a model for the entire class of demand-fetching memory policies that satisfy a resident-set inclusion property A generalized optimal policy (GOPT) is also defined; at its operating points it minimizes aggregated retention and swapping costs Special cases of the cost structure allow GWS and GOPT to simulate any known stack algorithm, the working set, and VMIN Efficient procedures for computing demand curves showing swapping load as a function of memory usage are developed for GWS and GOPT policies Empirical data from an actual system are included

/pdf/generalized-working-sets-for-segment-reference-strings-3u40e3vxq7.pdf

Generalized working sets for segment reference strings

Cold-start miss ratios are miss ratios that are measured with an initially empty first-level store. The values obtained depend on C, the first-level storage capacity, and on T, the number of references. These miss ratios, measured for various values of T, are useful in evaluating the effect of task switching on cache miss ratios when the cache capacity is C. For a particular (C, T) a practical method is described for obtaining the average of the w - T + 1 cold-start miss ratios measured on the substrings of length T of a reference string of length w. The method, which applies to set-associative LRU, can be used to find average cold-start miss ratios for a matrix of (C, T) pairs in one pass. In the general case, the computation time increases approximately logarithmically in the number of (C, T) pairs as the number of pairs is increased.

Computation of Cold-Start Miss Ratios

This paper describes the programming language facil ities required to simulate computer systems. The systems considered include large multiprocessor, virtual memory computing systems, and networks of these systems. The models considered range from those which only roughly describe system resources to those which fully describe all user and system software. The complex models at the latter level require powerful language facilities which promote program modularity. These basic language facilities include list processing and programmer-definable data structures and data types. The relevant features of the following languages are considered: the general- purpose languages FORTRAN, ALGOL 60, PL/I, ALGOL 68, SIMULA 67, and PASCAL; the general-purpose simulation languages SIMSCRIPT II, GPSS V, GASP, and SPURT 70; the computer simulation languages CSS, OSSL, ECSS, ASPOL, PSML, and CREAM; and the system implementation languages CONCURRENT PASCAL, CLU, ALPHARD, MODULA, and EUCLID. Of these languages only SIMULA 67...

Programming languages for computer system simulation

Transient-free average working-set size and transient-free missing-page rate for a finite sample of a reference string are defined. Use of these statistics is appropriate if the contents of the working set at the start of the recorded string are unknown. If a certain stationarity condition holds, these statistics provide unbiased estimates of expected working-set sizes, missing-page probabilities, and interreference distance probabilities. Two other pairs of estimators are shown to be biased. Expressions for the transient-free statistics are obtained in terms of interval statistics. Several methods of computation are discussed, the usefulness of each depending on length of the sample, number of distinct references, and the amount of main storage available to the computer performing the calculations. In particular, methods are described for handling long strings containing many distinct page names.

B. T. Bennett

Papers

Transient-free working-set statistics