A new tool for the performance analysis of massively parallel computer systems

We present a new tool, GPA, that can generate key performance measures for very large systems. Based on solving systems of ordinary differential equations (ODEs), this method of performance analysis is far more scalable than stochastic simulation. The GPA tool is the first to produce higher moment analysis from differential equation approximation, which is essential, in many cases, to obtain an accurate performance prediction. We identify so-called switch points as the source of error in the ODE approximation. We investigate the switch point behaviour in several large models and observe that as the scale of the model is increased, in general the ODE performance prediction improves in accuracy. In the case of the variance measure, we are able to justify theoretically that in the limit of model scale, the ODE approximation can be expected to tend to the actual variance of the model.

[1]  Peter G. Harrison,et al.  A probabilistic dynamic technique for the distributed generation of very large state spaces , 2000, Perform. Evaluation.

[2]  Allan Clark.  The ipclib PEPA Library , 2007, Fourth International Conference on the Quantitative Evaluation of Systems (QEST 2007).

[3]  C. Gillespie.  Moment-closure approximations for mass-action models. , 2009, IET systems biology.

[4]  Luca Cardelli,et al.  Efficient, Correct Simulation of Biological Processes in the Stochastic Pi-calculus , 2007, CMSB.

[5]  T. Kurtz.  Solutions of ordinary differential equations as limits of pure jump markov processes , 1970, Journal of Applied Probability.

[6]  Luca Cardelli,et al.  On process rate semantics , 2008, Theor. Comput. Sci..

[7]  Jane Hillston,et al.  A Compositional Approach to Performance Modelling (Distinguished Dissertations in Computer Science) , 2005 .

[8]  Mirco Tribastone.  The PEPA Plug-in Project , 2007, Fourth International Conference on the Quantitative Evaluation of Systems (QEST 2007).

[9]  William H. Sanders,et al.  Modelling techniques and tools for computer performance evaluation , 2006, Perform. Evaluation.

[10]  Jeremy T. Bradley,et al.  Evaluating fluid semantics for passive stochastic process algebra cooperation , 2010, Perform. Evaluation.

[11]  Jeremy T. Bradley,et al.  Spatial extension of stochastic Pi calculus , 2009 .

[12]  J. Bradley,et al.  A functional central limit theorem for PEPA , 2009 .

[13]  Marta Z. Kwiatkowska,et al.  PRISM: Probabilistic Symbolic Model Checker , 2002, Computer Performance Evaluation / TOOLS.

[14]  Stephen Gilmore,et al.  Derivation of passage-time densities in PEPA models using ipc: the imperial PEPA compiler , 2003, 11th IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer Telecommunications Systems, 2003. MASCOTS 2003..

[15]  Jeremy T. Bradley,et al.  A fluid analysis framework for a Markovian process algebra , 2010, Theor. Comput. Sci..

[16]  Chang Hyeong Lee,et al.  A moment closure method for stochastic reaction networks. , 2009, The Journal of chemical physics.

[17]  O. Kallenberg.  Foundations of Modern Probability , 2021, Probability Theory and Stochastic Modelling.

[18]  Hamid Bolouri,et al.  Dizzy: Stochastic Simulation of Large-scale Genetic Regulatory Networks , 2005, J. Bioinform. Comput. Biol..

[19]  Jane Hillston,et al.  Fluid flow approximation of PEPA models , 2005, Second International Conference on the Quantitative Evaluation of Systems (QEST'05).

[20]  Jane Hillston,et al.  A compositional approach to performance modelling , 1996 .

[21]  William H. Sanders,et al.  The Möbius Framework and Its Implementation , 2002, IEEE Trans. Software Eng..

[22]  Jeremy T. Bradley,et al.  Fluid passage-time calculation in large Markov models , 2009 .

[23]  Luca Cardelli,et al.  A Correct Abstract Machine for the Stochastic Pi-calculus , 2004 .

[24]  T. Kurtz.  Strong approximation theorems for density dependent Markov chains , 1978 .

[25]  D. Gillespie.  Exact Stochastic Simulation of Coupled Chemical Reactions , 1977 .

[26]  D. Gillespie.  Markov Processes: An Introduction for Physical Scientists , 1991 .

[27]  Alberto Policriti,et al.  Stochastic Concurrent Constraint Programming and Differential Equations , 2007, QAPL.