System-scenario-based design principles and applications

System-scenario-based design principles and applications / Francky Catthoor, Twan Basten, Nikolaos Zompakis, Marc Geilen, Per Gunnar Kjeldsberg.

This book introduces a generic and systematic design-time/run-time methodology for handling the dynamic nature of modern embedded systems, without adding large safety margins in the design. The techniques introduced can be utilized on top of most existing static mapping methodologies to deal effecti...

Full description

Saved in:
Bibliographic Details
Main Author: Catthoor, Francky
Other Authors: Basten, Twan, Zompakis, Nikolaos, Geilen, Marc, Kjeldsberg, Per Gunnar
Format: eBook
Language:English
Published: Cham : Springer, ©2020.
Subjects:
Embedded computer systems > Design and construction.
Embedded computer systems > Design and construction
Online Access:Click for online access
Table of Contents:
  • Intro; Preface; Contents; 1 Introduction and Organization of Book Material; 1 Motivation and Context; 2 Contributions of the Book and Target Audience; 3 Structure of the Book; 4 Classification; References; 2 System Scenario Methodology Flow; 1 Introduction and Context; 2 Use-Case Versus System Scenario Concept; 3 Motivating Example; 4 Basic Concepts and Terminology; 5 System Scenario Methodology; 5.1 Methodology Overview; 5.2 Identification; 5.2.1 RTS Parameter Discovery; 5.2.2 RTS Clustering; 5.3 Prediction; 5.4 Exploitation; 5.5 Switching; 5.6 Calibration; 6 Case Study Summary
  • 7 Extension to Multi-Tasking and Multi-Threading on Multi-Processor Platforms8 Related Work; 8.1 Related Design Approaches; 8.2 Scenario Exploitation Examples in Literature; 9 Conclusions; References; 3 System-Scenario-based Design Techniques in the Presence of Data Variables; 1 Introduction and Context; 2 Scenario Identification Through Polyhedral Partitioning of the Parameter Space; 2.1 Scenario Cost Definition for Use in Polyhedral Partitioning; 2.2 Algorithm for Polyhedral Scenario Identification; 2.3 Experimental Evaluation of Algorithm for Polyhedral Scenario Identification
  • 3 Scenario Identification Based on Specific Cost Parameters3.1 RTS Clustering Based on Memory Size and Frequency of Occurrence; 3.2 Clustering of RTSs Based on Image Size and Set of Available Platform Configuration Knobs; 4 Scenario Detection; 4.1 Scenario Prediction Using Application Monitoring Unit; 4.2 Scenario Prediction Through Precomputation; 5 Scenario Switching; 5.1 Scenario Switching Using Platform Adaptation Manager; 5.2 Switching Gain Evaluation; 6 Large-Scale Application Demonstrator; 6.1 Application, Platform, and Scenario System Settings; 6.2 Discussion of Obtained Results
  • 7 ConclusionsReferences; 4 DVFS-oriented Scenario Applications to Processor Architectures; 1 Software-Oriented Applications; 2 DVFS-RTH Sleep Mode Extensions; 2.1 Sleep Mode Management; 2.2 Sleep Mode Experimental Results; 3 Reliability-Sensitive Hardware-Oriented Applications and Gas-Pedal Extension; 3.1 Performance Dependability; 3.2 Introducing Gas-Pedal Points; 3.3 Choosing the Operating Points; 3.4 Case-Study Experiments; 3.4.1 Dependability in the Presence of Rollback Interventions; 3.4.2 Dependability in the Presence of Extra Load; 3.5 Hardware-Related Limitations of Our Scheme
  • 4 ConclusionsReferences; 5 DVAFS-Dynamic-Voltage-Accuracy- Frequency-Scaling Applied to Scalable Convolutional Neural Network acceleration; 1 Exploiting Dynamic Precision Requirements in DVAFS; 1.1 DAS: Dynamic-Accuracy-Scaling; 1.2 DVAS: Dynamic-Voltage-Accuracy-Scaling; 1.3 DVAFS: Dynamic-Voltage-Accuracy-Frequency-Scaling; 2 DVAFS Performance Analysis; 2.1 Performance of a DVAFS Multiplier; 2.2 Performance of a DVAFS SIMD Processor; 3 A DVAFS Prototype; 3.1 Envision: A DVAFS-Compatible CNN Processor; 3.2 Envision in a Face Recognition Hierarchy; 4 DVAFS Overview; References

Similar Items