Performance Models of Multiprocessor SystemsWhile there are several studies of computer systems modeling and performance evaluation where models of multiprocessor systems can be found as examples of applications of general modeling techniques, this is the first to focus entirely on the problem of modeling and performance evaluation of multiprocessor systems using analytical methods. Increasingly sophisticated and fast-moving technologies require models that can estimate the performance of a computer system without having actually to build and test it, models that can help designers make the correct architectural choices. The area of distributed computer architectures, or multiprocessor systems, has numerous such choices and can greatly benefit from an extensive use of performance evaluation techniques in the system design stage. The multiprocessor features that are studied here focus on contention for physical system resources, such as shared devices and interconnection networks. A brief overview covers the modeling of other important system characteristics, such as failures of components and synchronizations at the software level. Contents:Stochastic Processes. Queuing Models. Stochastic Petri Nets. Multiprocessor Architectures. Analysis of Crossbar Multiprocessor Architecture. Single Bus Multiprocessors with External Common Memory. Multiple Bus Multiprocessors with External Common Memory. Single Bus Multiprocessors with Distributed Common Memory. Multiple Bus Multiprocessors with Distributed Common Memory. The authors are affiliated with Politecnico and Universita di Torino in Italy. Performance Models of Multiprocessor Systemsis included in The MIT Press Series in Computer Systems, edited by Herb Schwetman. |
Other editions - View all
Performance Models of Multiprocessor Systems M. Ajmone Marsan,Gianfranco Balbo,Gianni Conte No preview available - 1986 |
Common terms and phrases
access requests analysis approximate assume average number balance equations BCMP theorem behavior buses busy memory modules common memory modules complexity comprises computing systems condition considered continuous-time Markov chain Coxian distribution CTMC cycle defined denote distributed computing DTMC enabled ergodic evaluation exponentially distributed FCFS finite function geometrically distributed global bus GSPN in figure GSPN model immediate transitions input interconnection network M₁ macrostates Markov property memoryless multiprocessor system normalization constant number of busy number of common number of customers number of processors number of tokens obtained p₁ parameters passive resources Petri Nets Poisson Poisson process possible private memory probability distribution processing power product form solution queue queuing network random variables reachability set recurrent represent service time distributions shown in figure slot steady state distribution steady state probability Stochastic Petri stochastic process subsystem t₂ transition probability matrix transition rate diagram vector workload