System Synthesis for Embedded Multiprocessors

Modern embedded systems must increasingly accommodate dynamically changing operating environments, high computational requirements, and tight time-to-market windows. Such trends and the ever-increasing design complexity of embedded systems have challenged designers to raise the level of abstraction and replace traditional ad-hoc approaches with more efficient synthesis techniques


1 Introduction
1.1 Multiprocessor Embedded Systems
1.2 Embedded Systems Design Automation
1.3 Contributions of this Thesis
1.3.1 Two-step Embedded Multiprocessor Scheduling
1.3.2 Clustering-based Heterogeneous Multiprocessor Scheduling
1.3.3 Multi-mode multi-task Embedded Systems Synthesis
1.3.4 Combined Assignment, Scheduling and Power Management Techniques
1.4 Outline of Thesis
2 System Synthesis: Definitions and Assumptions
2.1 System-Level Synthesis
2.2 Complexity of the Synthesis Problem
2.2.1 Optimization Algorithms
2.2.2 Multi-objective Optimization
2.2.3 Multi-Objective Evolutionary Algorithms (MOEA) Optimization
2.3 System Specification
3 Efficient Techniques for Clustering-Oriented Scheduling onto Homogeneous Embedded Multiprocessors
3.1 Background
3.1.1 Clustering and Scheduling
3.1.2 Genetic Algorithms
3.1.3 Existing Approaches
3.2 The Proposed Mapping Algorithm and Solution Description
3.2.1 CFA:Clusterization Function Algorithm
3.2.2 Randomized Clustering : RDSC, RSIA
3.2.3 Merging
3.2.4 Two-phase mapping
3.2.5 Comparison Method
3.3 Input Benchmark Graphs
3.3.1 Referenced Graphs
3.3.2 Application Graphs
3.3.3 Random Graphs
3.4 Performance Evaluation and Comparison
3.4.1 Results for the Referenced Graphs (RG) Set
3.4.2 Results for the Application Graphs (AG) Set
3.4.3 Results for the Random Graphs (RANG) Set
3.5 Summary and Conclusions
4 CHESS: Clustering-Oriented Heuristics for Heterogeneous Systems Scheduling
4.1 Related Work
4.2 Problem Statement
4.3 CHESS: Our proposed solution
4.3.1 CHESS-SCDM: Separate Clustering and Deterministic Merging
4.3.2 CHESS-SCGM: Separate Clustering and GA-based Merging
4.3.3 CHESS-CCDM: Combined Clustering and Deterministic Merging
4.3.4 CHESS-CCGM: Combined Clustering and GA-based Merging
4.4 The Heterogeneous-Earliest-Finish-Time (HEFT) Algorithm
4.4.1 The Randomized HEFT (RHEFT) Algorithm
4.5 Input Benchmark Graphs
4.6 Experimental Results
4.6.1 Performance study with respect to computation cost estimates
4.6.2 Performance study of different heterogeneous scheduling algorithms
4.7 Summary and Conclusions
5 CHARMED: A Multi-objective Co-synthesis Framework for Multi-mode Embedded Systems
5.1 Related Work
5.2 Problem statement
5.3 Evolutionary Multi-objective Optimization
5.4 CHARMED: Our Proposed Algorithm
5.4.1 MCFA: Multi-Mode Clusterization Function Algorithm
5.4.2 coreEA: mapping and scheduling
5.4.3 Multi-mode genetic operators
5.5 CHARMED-plus: Our Proposed Algorithm
5.6 Parallel CHARMED
5.7 Experimental results
5.8 Summary and Conclusions
6 CASPER: An Integrated Framework for Energy-Driven Scheduling on Embedded Multiprocessor Systems
6.1 Problem Statement and Assumptions
6.2 Proposed Algorithmic Solution
6.2.1 Combined Assignment and Scheduling
6.2.2 Power Management Techniques
6.2.3 Refinement
6.3 Experimental Results
6.3.1 Homogeneous System
6.3.2 Heterogeneous System
6.4 Conclusions
7 Conclusions and Future Work

Author: Kianzad, Vida

Source: University of Maryland

Download URL 2: Visit Now

Leave a Comment