Trace-driven simulation is a popular technique useful in many applications, as for example analysis of memory hierarchies or internet subsystems, and to evaluate the performance of computer systems. Normally, traces should be gathered from really working systems. However, real traces require enormous memory space and time. An alternative is to generate Synthetic traces using suitable algorithms. In this paper we describe an algorithm for the synthetic generation of memory references which behave as those generated by given running programs. Our approach is based on a novel Machine Learning algorithm we called Hierarchical Hidden/non Hidden Markov Model (HHnHMM). Short chunks of memory references from a running program are classified as Sequential, Periodic, Random, Jump or Other. Such execution classes are used to train an HHnHMM for that program. Trained HHnHMM are used as stochastic generators of memory reference addresses. In this way we can generate in real time memory reference streams of any length, wich mimic the behaviour of given programs without the need to store anything. It is worth noting that our approach can be extended to other applications, for example network or data storage systems. In this paper we address only the generation of synthetic memory references generated by instruction fetches. Experimental results and a case study conclude this paper.

Towards Effective Generation of Synthetic Memory References Via Markovian Models

Cuzzocrea, Alfredo;
2018

Abstract

Trace-driven simulation is a popular technique useful in many applications, as for example analysis of memory hierarchies or internet subsystems, and to evaluate the performance of computer systems. Normally, traces should be gathered from really working systems. However, real traces require enormous memory space and time. An alternative is to generate Synthetic traces using suitable algorithms. In this paper we describe an algorithm for the synthetic generation of memory references which behave as those generated by given running programs. Our approach is based on a novel Machine Learning algorithm we called Hierarchical Hidden/non Hidden Markov Model (HHnHMM). Short chunks of memory references from a running program are classified as Sequential, Periodic, Random, Jump or Other. Such execution classes are used to train an HHnHMM for that program. Trained HHnHMM are used as stochastic generators of memory reference addresses. In this way we can generate in real time memory reference streams of any length, wich mimic the behaviour of given programs without the need to store anything. It is worth noting that our approach can be extended to other applications, for example network or data storage systems. In this paper we address only the generation of synthetic memory references generated by instruction fetches. Experimental results and a case study conclude this paper.
978-1-5386-2666-5
Trace-driven simulation
ergodic HMM
memory references
execution classes
synthetic memory references
spectral analysis.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/20.500.11770/312764
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? ND
social impact