{"title":"Discrete simulation of core utilization","authors":"Edgar Martin, L. J. Taylor","doi":"10.1145/503561.503586","DOIUrl":null,"url":null,"abstract":"The purpose of this simulation is to determine if the current computer hardware at the University of Mississippi is sufficient for projected needs and if by scheduling the use period of some users, more efficient service could be given to all users. All programs involved in the simulation are in Standard FORTRAN and are therefore easily adaptable to all computer facilities.By implementation of this simulation procedure, any computer facility can determine the efficiency of core utilization with their current scheduling procedures as compared to alternate systems. The scheduling of \"fixed\" or constant users of the system to a low use time period, can be simulated to determine if an overall increase in core utilization is possible. As an added feature this simulation procedure can forecast whether the current core capacity is sufficient for future needs. These two simulations can be tied together to allow a facility to determine how long current core capacity will be sufficient. Because these changes are simulated, increased core efficiency can be tested without disrupting normal routine. Same of the detailed results provided by this project are: Average amount of core available(under the simulated conditions); per cent of time core is full; amount of time a user must wait if core is full, number of jobs run for each time period; and statistics on the individual jobs run during the time period.The programs involved in the simulation function as follows. Jobs are placed in a core input queue at random time intervals, based on observed conditions. The jobs are then loaded in core as space permits and removed at completion times. Each job's core requirement, time in core, and origin is based on actual data but are also selected by random procedures to insure realizm.The simulation of actual conditions at the University of Mississippi proved to be extremely accurate. The results of simulations conducted have proved to be both plausible and useful.","PeriodicalId":151957,"journal":{"name":"ACM-SE 14","volume":"135 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1976-04-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACM-SE 14","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/503561.503586","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
The purpose of this simulation is to determine if the current computer hardware at the University of Mississippi is sufficient for projected needs and if by scheduling the use period of some users, more efficient service could be given to all users. All programs involved in the simulation are in Standard FORTRAN and are therefore easily adaptable to all computer facilities.By implementation of this simulation procedure, any computer facility can determine the efficiency of core utilization with their current scheduling procedures as compared to alternate systems. The scheduling of "fixed" or constant users of the system to a low use time period, can be simulated to determine if an overall increase in core utilization is possible. As an added feature this simulation procedure can forecast whether the current core capacity is sufficient for future needs. These two simulations can be tied together to allow a facility to determine how long current core capacity will be sufficient. Because these changes are simulated, increased core efficiency can be tested without disrupting normal routine. Same of the detailed results provided by this project are: Average amount of core available(under the simulated conditions); per cent of time core is full; amount of time a user must wait if core is full, number of jobs run for each time period; and statistics on the individual jobs run during the time period.The programs involved in the simulation function as follows. Jobs are placed in a core input queue at random time intervals, based on observed conditions. The jobs are then loaded in core as space permits and removed at completion times. Each job's core requirement, time in core, and origin is based on actual data but are also selected by random procedures to insure realizm.The simulation of actual conditions at the University of Mississippi proved to be extremely accurate. The results of simulations conducted have proved to be both plausible and useful.