Browse Prior Art Database

Algorithm for Determining Optimum Configurations of Memory Hierarchies

IP.com Disclosure Number: IPCOM000078400D
Original Publication Date: 1972-Dec-01
Included in the Prior Art Database: 2005-Feb-25
Document File: 2 page(s) / 53K

Publishing Venue

IBM

Related People

Chow, CK: AUTHOR

Abstract

With the algorithm disclosed herein, there is enabled the design of a memory hierarchy for a given total system cost S(o), and a system capacity C(n). More specifically, this algorithm enables the determining of the salient system parameters such as the number of memory levels, NOPT, the capacities C(i) at each level, and the access time, t(i) of each level. In this latter connection, the access time, t(i), determines the technology. The end which is sought is the minimization of the effective hierarchy access time, T, for each memory reference.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 62% of the total text.

Page 1 of 2

Algorithm for Determining Optimum Configurations of Memory Hierarchies

With the algorithm disclosed herein, there is enabled the design of a memory hierarchy for a given total system cost S(o), and a system capacity C(n). More specifically, this algorithm enables the determining of the salient system parameters such as the number of memory levels, NOPT, the capacities C(i) at each level, and the access time, t(i) of each level. In this latter connection, the access time, t(i), determines the technology. The end which is sought is the minimization of the effective hierarchy access time, T, for each memory reference.

In considering the algorithm, reference is made to Fig. 1 wherein there is shown a linear memory hierarchy which consists of n levels, viz., M(1), M(2),...,M(n) in cascade arrangement. Generally, the lesser the number, n, of levels. the faster the speed of the memory, the higher its cost per byte, and the smaller its capacity. Data transfers are between adjacent levels, each level M(i) is characterized by its access time, t(i) (in seconds) and capacity C(i) (bytes or some suitable unit determined by the constant F(0), as set forth in (3) hereinbelow). The cost per byte is dependent upon the access time which is conveniently referred to as b(t(i)).

The flow chart depicted in Fig. 2 sets forth the algorithm in relatively broad conceptual terms. In this flow chart, the values which are input as represented by input block 10 are; 1) the given system cost,...