liu.seSearch for publications in DiVA
Change search
ReferencesLink to record
Permanent link

Direct link
Improved Scheduling Techniques for Parallel Distributed-Solver System Simulation
Linköping University, Department of Management and Engineering, Fluid and Mechatronic Systems. Linköping University, Faculty of Science & Engineering.
Linköping University, Department of Management and Engineering, Fluid and Mechatronic Systems. Linköping University, Faculty of Science & Engineering.
Linköping University, Department of Management and Engineering, Fluid and Mechatronic Systems. Linköping University, Faculty of Science & Engineering.ORCID iD: 0000-0002-2315-0680
(English)Manuscript (preprint) (Other academic)
Abstract [en]

Shortening simulation time is an important step towards efficient simulation-based product development. A long-used method is to exploit physically motivated time delays to split up the model into distributed solvers. In this way, the use of a centralized sequential solver can be circumvented. For maximum simulation performance, however, an efficient scheduling technique is also required. Four task scheduling methods for distributed-solver simulations has been implemented and evaluated. Experiments indicate that the best choice largely depend on model size, load distribution and granularity. Lock-based barrier synchronization provides the highest speed-up for small models. A fork-join implementation, with implicit synchronization and work-stealing scheduling, works better for models with a large total workload. It is common that workload and load distribution of a simulation model varies during execution depending on the current state of the simulation. Three of the implemented schedulers support dynamic load balancing during execution. Results show that task-stealing is the most efficient method for the specific test model. A possible continuation of this work is an automatic selection of the best scheduling technique based on knowledge about model properties and available computer resources.

Keyword [en]
System simulation, distributed solvers, parallelism, scheduling, transmission line element method
National Category
Electrical Engineering, Electronic Engineering, Information Engineering Fluid Mechanics and Acoustics
Identifiers
URN: urn:nbn:se:liu:diva-122749OAI: oai:DiVA.org:liu-122749DiVA: diva2:872635
Available from: 2015-11-19 Created: 2015-11-19 Last updated: 2015-11-19
In thesis
1. Distributed System Simulation Methods: For Model-Based Product Development
Open this publication in new window or tab >>Distributed System Simulation Methods: For Model-Based Product Development
2015 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Distributed system simulation can increase performance, re-usability and modularity in model-based product development. This thesis investigates four aspects of distributed simulation: multi-threaded simulations, simulation tool coupling, distributed equation solvers and parallel optimization algorithms.

Multi-threaded simulation makes it possible to split up the workload over several processing units. This reduces simulation time, which can save both time and money during the product development cycle. The transmission line element method (TLM) is used to decouple models to independent sub-models.

Different simulation tools are suitable for different problems. Tool coupling makes it possible to use the best suited tool for simulating each part of the whole product. Models from different tools can then be coupled into one aggregated simulation model. An emerging standard for tool coupling is the Functional Mock-up Interface (FMI). It is investigated how this can be used in conjunction with TLM.

Equation-based object-oriented languages (EOOs) are becoming increasing popular. A logical approach is to let the equation solvers maintain the same structure that was used in the modelling process. Methods for achieving this using TLM and FMI are implemented and evaluated.

In addition to parallel simulations, it is also possible to use parallel optimization algorithms. This introduces parallelism on several levels. For this reason, an algorithm for profile-based multi-level scheduling is proposed.

Place, publisher, year, edition, pages
Linköping: Linköping University Electronic Press, 2015. 118 p.
Series
Linköping Studies in Science and Technology. Dissertations, ISSN 0345-7524 ; 1732
National Category
Electrical Engineering, Electronic Engineering, Information Engineering Fluid Mechanics and Acoustics
Identifiers
urn:nbn:se:liu:diva-122754 (URN)10.3384/diss.diva-122754 (DOI)978-91-7685-875-2 (print) (ISBN)
Public defence
2015-12-18, ACAS, A-huset, Campus Valla, Linköping, 10:15 (English)
Opponent
Supervisors
Available from: 2015-11-19 Created: 2015-11-19 Last updated: 2015-12-02Bibliographically approved

Open Access in DiVA

No full text

Search in DiVA

By author/editor
Braun, RobertNordin, PeterKrus, Petter
By organisation
Fluid and Mechatronic SystemsFaculty of Science & Engineering
Electrical Engineering, Electronic Engineering, Information EngineeringFluid Mechanics and Acoustics

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 267 hits
ReferencesLink to record
Permanent link

Direct link