G. Balbo, S. Donatelli, and G. Franceschinis, "Understanding Parallel Program Behavior through Petri Net Models," Journal of Parallel and Distributed Computing, Vol. 15, pp.171-187, 1992. J. Zhao et al. / Run-Time Detection of Communication Deadlocks 107  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the other hand, dynamic detection methods monitor the behavior of a concurrent program at run time and detect deadlocks in the program without reporting nonexistent deadlocks. However, although a number of methods have been proposed for static detection of communication deadlocks in occam programs [1, 7, 8, 10, 11, 12], no method has been proposed for dynamic detection of communication deadlocks in occam programs. In this paper we propose a novel approach to detect communication deadlocks in occam 2 programs at run time. The main feature of our approach is to monitor the interprocess communications of a target ....

....was based on that developed by Xu et al. 12] but has been extensively modified and expanded. The algorithm is a static checker, thus, it is inevitable for it to report some nonexistent deadlocks sometimes during the period of checking. In addition to [12, 13, 10] there are some other papers [1, 7, 8, 11] concerning the deadlock detection of occam programs. These methods are all based on the static analysis of the Petri net model of an occam program just like introduced above, and therefore suffer the same problem with [12, 13, 10] such that may report nonexistent deadlocks. 7 Conclusions We ....

G. Balbo, S. Donatelli, and G. Franceschinis, "Understanding Parallel Program Behavior through Petri Net Models," Journal of Parallel and Distributed Computing, Vol. 15, pp.171-187, 1992. J. Zhao et al. / Run-Time Detection of Communication Deadlocks 107

....between PNs and concurrent languages like CSP and CCS (Occam2 actually uses CSP models [Hoare85] Instead, Taylor [Taylor83] and Shatz and Cheng [Shatz87] consider Ada. Taylor proposes a static analyzer for Ada programs. It inputs the program code and outputs the program call graph. Balbo et al. Balbo92a] report part of a project for the development of a Petri net based programming environment for the DISC programming language that includes the CSP parallel constructs. Although Petri nets have many good properties which make them superior to other formalisms (at least at some important aspects) ....

....variable (which obeys an exponential distribution) The transition whose timer goes to zero first is the one selected to fire. The dynamic operation of a GSPN is equivalent to a continuous time stochastic process. An example of using GSPNs in parallel program performance study can be found in [Balbo92a] DSPNs (Deterministic and Stochastic Petri Nets) described in [Ajmone Marsan87] are another class of SPNs, which allow both deterministic and random firing times, and thus are useful for many practical situations. The analysis tool has also been published [Lindemann94] An application can be ....

Balbo, G., Donatelli, S. and Franceschinis, G., "Understanding parallel program behavior through Petri net models," Journal of Parallel and Distributed Computing, vol. 15, pp. 171-187, 1992.

....However, they are less useful for describing the logical behaviour of a hierarchical SW system with fork join effects and refinement. Petri nets are a more prominent formalism for this purpose, they are suitable to model typical language constructs of imperative, concurrent programming languages [3]. Incorporating a stochastic notion of time into Petri nets yields the well known generalised stochastic Petri nets (GSPNs) However GSPNs do not show sufficient support for a structured way of modeling, such that design methodologies have been developed to demonstrate how to use GSPNs for a ....

....can be neatly composed into a hierarchical structure. Advantages of a hierarchical description for validation and performance analysis are briefly sketched in Sect. 5. A summary is given in Sect. 6. 2 Specification of Software with HQPNs Modeling program flow with Petri nets is well known [3] and with the introduction of colours also data dependent behaviour can be encoded with moderate effort [19, 20] For small programs the modeling process is thus straight forward. The development of larger programs, or SW systems in general, is normally a top down process and requires additional ....

G. Balbo, S. Donatelli, J. Franceschinis; Understanding parallel program behavior through Petri net models; J. Parallel and Distributed Comp. 15 (3) 1992, 171-187.

....K. Nichols [93] ffl Performance Measurements Tools in a Multiprocessor Environment, H. Burkhart and B. Miller [94] ffl A Petri net approach for performance oriented parallel program design, A. Ferscha [95] ffl Understanding parallel program behavior through Petri net models, G. Balbo et al. [96]. ffl A methodology for performance analysis of parallel computations with looping constructs, A. Kapelnikov et al. [97] ffl A methodology for performance evaluation of parallel applications on multiprocessors, D. Menasce and L. Barroso [98] ffl Timing parallel programs that use message ....

G. Balbo, S. Donatelli, and G. Franceschinis, "Understanding parallel program behavior through petri net models", Journal of Parallel and Distributed Computing, vol. 15, no. 3, pp. 171--187, 1992.

....discusses a more advanced estimation approach using symbolic analysis. A. van Gemund ( 17] designed the Pamela performance modeling methodology, which provides a theoretical framework for modeling and analyzing serialization effects and the performance of parallel systems. Although petri nets ([2, 15]) queueing networks ( 22] and markov chains ( 30] can be valuable in understanding the dynamic behavior of parallel programs, the associated analysis costs both in terms of runtime and memory requirements prohibits their use in compilers. N. Yazici Pekergin and J.M. Vincent ( 33] obtain ....

G. Balbo, S. Donatelli, and G. Franceschinis. Understanding parallel program behavior through petri net models. Journal of Parallel and Distributed Computing, 15:181--187, 1992.

....[Tay83b] imply that, in general, the time and space requirements for this technique are exponential. Methods proposed for improving accuracy of reachability analysis include combining reachability analysis with symbolic execution [YT88] and using program variable value information in the analysis [BDF92] The accuracy improving techniques in the following sections are presented in the context of reachability analysis based on a Petri net representation of the programs to be analyzed. Symbolic model checking techniques [BCM 90] represent the program state space symbolically rather than ....

....value place for each possible value of the variable, plus an Unknown place to account for those occasions on which we can not statically determine the variable value. The variable subnet for a Boolean variable would thus have a True place, a False place, and an Unknown place. Balbo et al. BDF92] suggest an alternate approach in which each variable has a single place, with the variable s value represented by the number of tokens in the place; we choose not to use this approach because it yields an unsafe net. The variable subnet that we use also includes operation places for the valid ....

Gianfranco Balbo, Susanna Donatelli, and Giuliana Franceschinis. Understanding parallel program behavior through petri net models. Journal of Parallel and Distributed Computing, 15(3):171-187, July 1992.

No context found.

G. Balbo, S. Donatelli, and G. Franceschinis. "Understanding Parallel Program Behavior through Petri Net Models". Journal of Parallel and Distributed Computing, Vol. 15, No. 3, pp. 171--187, July 1992.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC