Analysis of a Drum Input/Output Queue Under Scheduled Operation in a Paged Computer System
Journal of the ACM (JACM)
A Survey of Analytical Time-Sharing Models
ACM Computing Surveys (CSUR)
Dynamic space-sharing in computer systems
Communications of the ACM
Dynamic storage allocation systems
Communications of the ACM
The working set model for program behavior
Communications of the ACM
Dynamic program behavior under paging
ACM '66 Proceedings of the 1966 21st national conference
Program behavior and control in virtual storage computer systems
Program behavior and control in virtual storage computer systems
Experience using a time-shared multi-programming system with dynamic address relocation hardware
AFIPS '67 (Spring) Proceedings of the April 18-20, 1967, spring joint computer conference
Scheduling TSS/360 for responsiveness
AFIPS '70 (Fall) Proceedings of the November 17-19, 1970, fall joint computer conference
The page fault frequency replacement algorithm
AFIPS '72 (Fall, part I) Proceedings of the December 5-7, 1972, fall joint computer conference, part I
Multiprocessor Scheduling with Memory Allocation A Deterministic Approach
IEEE Transactions on Computers
| Hi-index | 0.01 |
In a multiprogramming, virtual-memory computing system, many processes compete for the main storage page frames and CPU's of the real system. It is customary to define a subset of these processes called the "multiprogramming set" (MPS), and to allocate resources only to those processes currently in the MPS. Each process remains in the MPS for a limited time and is then demoted. The system paging manager controls the size of the MPS; it allocates the available page frames among the processes in the MPS and fetches appropriate pages into the page frames. A model is described that assumes the most critical resources of the system to be page frames and the paging channel (i.e., there is no significant CPU contention). The model makes certain assumptions about the page fault rate of processes as a function of page frames allocated, and about the page fetch time as a function of mean load on the paging channel. The model also incorporates a definition of the value of a given page allocation in terms of system throughput. The model is used to study various strategies for choosing an MPS and allocating page frames among processes. For simple cases, the model yields an exact optimal strategy. A heuristic strategy is proposed for dealing with more complex cases, and is shown by the model to be reasonably near optimal. The heuristic strategy monitors the page fault rate of each process and chooses an allocation such that each process can be executed at a reasonable rate, while ensuring that the paging channel is neither overloaded nor underloaded.