Frequency interleaving as a codesign scheduling paradigm

  • Authors:

  • JoAnn M. Paul;Simon N. Peffers;Donald E. Thomas

  • Affiliations:

  • Center for Electronic Design Automation, Carnegie Mellon University, Pittsburgh, PA;Center for Electronic Design Automation, Carnegie Mellon University, Pittsburgh, PA;Center for Electronic Design Automation, Carnegie Mellon University, Pittsburgh, PA

  • Venue:
  • CODES '00 Proceedings of the eighth international workshop on Hardware/software codesign
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

Frequency interleaving is introduced as a means of conceptualizing and co-scheduling hardware and software behaviors so that software models with conceptually unbounded state and execution time are resolved with hardware resources. The novel mechanisms that result in frequency interleaving are a shared memory foundation for all system modeling (from gates to software-intensive subsystems) and de-coupled, but interrelated time- and state-interleaved scheduling domains. The result for system modeling is greater accommodation of software as a configuration paradigm that loads system resources, a greater accommodation of shared memory modeling, and a greater representation of software schedulers as a system architectural abstraction. The results for system co-simulation are a lessening of the dependence on discrete event simulation as a means of merging physical and non-physical models of computation, and a lessening of the need to partition a system as computation and communication too early in the design. We include an example demonstrating its implementation.