Issues in multiprogrammed multiprocessor scheduling
We also compare the policies under workloads that may be more realistic than previous studies have used. Using these new workloads, we arrive at different conclusions than reported in earlier work. In particular, we find that the 'smallest number of processes first' (SNPF) scheduling discipline performs poorly, even when the number of processes in a job is positively correlated with the total service demand of the job. We also find that policies that allocate an equal fraction of the processing power to each job in the system perform better than practical policies that allocate processing power unequally. We find that allocation of processing power among competing jobs is at least as important as explicit support for spin-lock and barrier synchronization.
15 pages matching open system in this book
Results 1-3 of 15
What people are saying - Write a review
We haven't found any reviews in the usual places.
Multiprocessor and Workload Models
4 other sections not shown
allocate processing power allocation of processing allocation per job application level scheduling barrier synchronization baseline cache misses chapter Closed System Uncorrelated Cn d cd coefficient of variation consider context switch correlated geometric-bounded workload correlated workload Cosched cv t Input deschedule processes EqualDP and UnequalDP EqualDP performs exponential distribution Figure fraction of processing geometric distribution hyperexponential workload increases input parameter job lock demand load queue mean job demand mean job parallelism mean job response multi-level queue multiprocessor scheduling policies number of processes open system Output Percent Preemption Overhead plots mean response policies FCFS policies that allocate preempt processes preemption frequency process cache processes holding processes per job provide equal allocation PSCDF PSNPF reach the barrier reaching a barrier Response Time Ratios response time versus RRjob RRprocess Shortest Remaining slice smallest number SNPF spin-lock synchronization spinning support for inter-process thesis tion uncorrelated geometric-bounded uniprocessor scheduling University of Wisconsin-Madison variation of job