Performance Analysis Of Devs Environments
Ezequiel Glinsky, Gabriel Wainer, Planta Baja Pabellón I, Ciudad Universitaria
Abstract
Ezequiel Glinsky, Gabriel Wainer, Planta Baja Pabellón I, Ciudad Universitaria
Abstract
The CD++ toolkit was developed in order to implement the theoretical concepts specified by the DEVS formalism. CD++ has been enhanced lately to support both parallel and real-time simulation. The abstract simulation algorithms associated with the hierarchical modelling techniques have a cost in terms of processing time overhead. Besides this, using the tool for real-time modelling involve task models that usually do not take overhead into account. Our goal is to characterize the overhead of stand-alone, parallel and real-time simulators available in the toolkit. To do so, we used a synthetic benchmarking tool that can be applied to DEVS environments. Different types of models are tested automatically, making the performance analysis easy. The results here presented show that we are able to develop DEVS models paying a small cost in terms of processing overhead.
OpenAlex reports 4 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The CD++ toolkit was developed in order to implement the theoretical concepts specified by the DEVS formalism. CD++ has been enhanced lately to support both parallel and real-time simulation. The abstract simulation algorithms associated with the hierarchical modelling techniques have a cost in terms of processing time overhead. Besides this, using the tool for real-time modelling involve task models that usually do not take overhead into account. Our goal is to characterize the overhead of stand-alone, parallel and real-time simulators available in the toolkit. To do so, we used a synthetic benchmarking tool that can be applied to DEVS environments. Different types of models are tested automatically, making the performance analysis easy. The results here presented show that we are able to develop DEVS models paying a small cost in terms of processing overhead.
Key concepts: DEVS, Computer science, Simulation, Modeling and simulation