Messung und Modellierung paralleler und verteilter Rechensysteme

Klar, Rainer; Dauphin, Peter; Hartleb, Franz; Hofmann, Richard; Mohr, Bernd; Quick, Andreas; Siegle, Markus

doi:10.1007/978-3-322-99681-7

Cited by 19 publications

(5 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If a task graph is not series-parallel reducible it can be analysed using several well-known approximate (bounding) methods [15,21]. The basic idea behind these approximations is that nodes are added or deleted until the task graph becomes series-parallel reducible.…”

Section: Task Graph Analysis and Peppmentioning

confidence: 99%

“…For series-parallel graphs numerical results are exact and algorithms exist to compute the distribution of the delay between a start and finish task. For arbitrary graphs various approximate techniques exist to compute (rather exact) bounds on the mean delay [21]. We use the Pepp tool suite [8,15] to analyse the task graphs generated from the complete prefixes.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

First Passage Time Analysis of Stochastic Process Algebra Using Partial Orders

Ruys

Langerak

Katoen

et al. 2001

Tools and Algorithms for the Construction and Analysis of Systems

View full text Add to dashboard Cite

This paper proposes a partial-order semantics for a stochastic process algebra that supports general (non-memoryless) distributions and combines this with an approach to numerically analyse the first passage time of an event. Based on an adaptation of McMillan's complete finite prefix approach tailored to event structures and process algebra, finite representations are obtained for recursive processes. The behaviour between two events is now captured by a partial order that is mapped on a stochastic task graph, a structure amenable to numerical analysis. Our approach is supported by the (new) tool Forest for generating the complete prefix and the (existing) tool Pepp for analysing the generated task graph. As a case study, the delay of the first resolution in the root contention phase of the IEEE 1394 serial bus protocol is analysed.

show abstract

Section: Task Graph Analysis and Peppmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

First Passage Time Analysis of Stochastic Process Algebra Using Partial Orders

Ruys

Langerak

Katoen

et al. 2001

Tools and Algorithms for the Construction and Analysis of Systems

View full text Add to dashboard Cite

show abstract

“…Furthermore, for requirements and demands of trace analysis software tools [7,8,13,17,18] are referenced.…”

Section: Related Workmentioning

confidence: 99%

A New Data Compression Technique for Event Based Program Traces

Knüpfer

2003

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The paper presents an innovative solution to the problem of the very huge data sets that are regularly produced by performance tracing techniques -especially on HPC programs. It designs an adapted data compression scheme that takes advantage of regularities frequently found in program traces. Algorithms to reveal repetition patterns in a programs call structure and run time behavior are discussed in detail, solutions to some problems arising on practical application are addressed as well. Two examples demonstrate the capabilities of the approach and document its behavior. Finally, some thoughts are given regarding how the patterns revealed in the process of data compression may assist the automatic analysis of traces.

show abstract

“…For series-parallel graphs numerical results are exact and algorithms exist to compute the distribution of the delay between a start and finish task. For arbitrary graphs various approximate techniques exist to compute (rather exact) bounds on the average delay [115]. We use the PEPP tool suite [45,70] to analyse the task graphs generated from the complete prefixes.…”

Section: Introductionmentioning

confidence: 99%

“…If a task graph is not series-parallel reducible it can be analysed using several well-known approximate (bounding) methods [70,115]. The basic idea behind these approximations is that nodes are added or deleted until the task graph becomes series-parallel reducible.…”

mentioning

confidence: 99%

Towards effective model checking

Ruijs¹

View full text Add to dashboard Cite

iii many years to come. I would like to thank one person specifically; the assistance of Joke Lammerink was invaluable in times of increased stress.Special thanks should go out to Pedro D'Argenio, who was my roommate for more than four and a half years. Although we had different opinions on several aspects of scientific life, e.g., tidiness, deadlines, theory or practice, students, education, etc., it kept the discussions lively and in all these years, we never had a quarrel. I really enjoyed his company and would like to thank him for all things he 'learned' me. At the same time, I should apologise for never being there when my telephone rang or when students came to pay me a visit. I regret it that Pedro will leave our group this summer, but I am sure that our friendship will survive the move to his home country.I would also like to thank Lex Heerink, with whom Pedro and I shared a room in the beginning of my Ph.D. student-ship. Among many other things, Lex made me enthusiastic for the IPA research school: not only for the contribution of this school to our Ph.D. education, but also for the 'social' significance of the school. Lex strongly believes that it is unhealthy to work all day. Since he left the group, the social 'bakje koffie' events have become very irregular. But (who's butt?) many of his running jokes are still being heard.Although I have never met him, I would like to thank René van Oostrum, who graciously sent me the L A T E X 2 ε sources of his Ph.D. Thesis. These sources contained a wealth of formatting tricks that I used to beautify the layout of this dissertation.I would like to thank my (little) brother Eric and Erik-Jan Elfferich for accepting the dubious honour of being my 'paranymphs'. I am confident that they will make a great performance on D-day.All work and no play makes Jack a dull boy. I would like to thank Anja Beekhuis, Ellen Tiemens and Joke de Bruin. Although not contributing directly to this dissertation, our bowling evenings on Monday were really very effective to help me relax and forget the thesis for a moment. In the same 'league', I acknowledge my year in the board of the LAIOO; I would like to thank Jeanette Simon, Femmie de Vegt and Rob Roovers for the enjoyable, constructive and laughable meetings.All scientific help and support from my colleagues would probably have been useless, if I would not have a 'home base' to fall back to. Mischa, I would like to thank you for all your patience, consideration, support and love during the many months that I needed to finish this thesis. Many times it was very hard to return to this dissertation, while you and our little Daphne were so nearby to have fun with. I promise that I will make up for all the times that I was not there.

show abstract

Messung und Modellierung paralleler und verteilter Rechensysteme

Cited by 19 publications

References 0 publications

First Passage Time Analysis of Stochastic Process Algebra Using Partial Orders

First Passage Time Analysis of Stochastic Process Algebra Using Partial Orders

A New Data Compression Technique for Event Based Program Traces

Towards effective model checking

Contact Info

Product

Resources

About