PSML: parallel system modeling and simulation language for electronic system level

System-level description languages in electronic system level are domain-specific sequential simulation languages using the shared-everything model in discrete-event modeling and simulation terminology. They implement sequential process-interaction worldview to take advantage of event partitioning f...

Full description

Saved in:
Bibliographic Details
Published in:The Journal of supercomputing 2019-05, Vol.75 (5), p.2691-2724
Main Authors: Poshtkohi, Alireza, Ghaznavi-Ghoushchi, M. B., Saghafi, Kamyar
Format: Article
Language:English
Subjects:
Compilers
Complexity
Computer Science
Computer simulation
Cyber-physical systems
Discrete event systems
Embedded systems
Flight simulators
Interpreters
Modelling
Modularity
Partitioning
Processor Architectures
Programming Languages
Simulation
Specification and description languages
Systems design
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:System-level description languages in electronic system level are domain-specific sequential simulation languages using the shared-everything model in discrete-event modeling and simulation terminology. They implement sequential process-interaction worldview to take advantage of event partitioning for ease of programming and modularity. Therefore, their reference simulators can only be executed on a single physical core. Fast and accurate simulation is highly desirable for efficient and effective system design due to the ever-increasing complexity of embedded and cyber physical systems. Parallel discrete-event simulation (PDES) is the main technique to solve this problem for large-scale system-level models. PDES works based on state space partitioning by using the so-called logical process worldview. This paper proposes parallel system modeling and simulation language (PSML), along with its formalized distributed parallel simulation kernel, that provides execution of hardware models in order to improve simulation speed significantly. It will be shown that the proposed framework results in linear, super-linear speedups ranging from 11 × to 32 × for large-scale, complex PSML models in comparison with the SystemC reference simulator on a 12-core host.
ISSN:0920-8542
1573-0484
DOI:10.1007/s11227-018-2682-1