Hardware software partitioning of task graph using genetic algorithm

2014 ◽  
Author(s):  
Ashish Mishra ◽  
Dhruv Vakharia ◽  
Anirban Jyoti Hati ◽  
Kota Solomon Raju
Keyword(s):  
Genetic Algorithm ◽  
Task Graph ◽  
Software Partitioning

A Multi-Objective Optimization Genetic Algorithm for SOPC Hardware-Software Partitioning

2012 ◽  
Vol 457-458 ◽  
pp. 1142-1148
Author(s):  
Fu Yang ◽  
Liu Xin ◽  
Pei Yuan Guo
Keyword(s):  
Genetic Algorithm ◽  
Optimization Problem ◽  
Search Algorithm ◽  
Multi Objective Optimization ◽  
Key Technology ◽  
Multi Objective ◽  
Multi Objective Genetic Algorithm ◽  
Partitioning Problem ◽  
System On Programmable Chip ◽  
Software Partitioning

Hardware-software partitioning is the key technology in hardware-software co-design; the results will determine the design of system directly. Genetic algorithm is a classical search algorithm for solving such combinatorial optimization problem. A Multi-objective genetic algorithm for hardware-software partitioning is presented in this paper. This method can give consideration to both system performance and indicators such as time, power, area and cost, and achieve multi-objective optimization in system on programmable chip (SOPC). Simulation results show that the method can solve the SOPC hardware-software partitioning problem effectively.

Research on Hardware-Software Partitioning Method for Dynamically Reconfigurable Embedded System

2012 ◽  
Vol 198-199 ◽  
pp. 606-610
Author(s):  
Xin Zhao ◽  
Lan Ying Li ◽  
Ming Gao
Keyword(s):  
Embedded System ◽  
System Model ◽  
Ant Algorithm ◽  
Task Graph ◽  
List Scheduling ◽  
Dynamically Reconfigurable ◽  
System A ◽  
Application Model ◽  
Reconfigurable System ◽  
Software Partitioning

According to the need of reconfigurable system’s designing, this paper presents a method applied to hardware-software partitioning for dynamically reconfigurable embedded system. A model of dynamically reconfigurable embedded system is built to describe its task; then a merging algorithm of genetic algorithm and ant algorithm is adopt to get results of hardware-software partitioning, according to the system model given and application model described by task graph. Priority-based list scheduling and Configuration-prefetching strategy are used to evaluate these results. Experimental results show that the algorithm in this paper could map the task graphs to reconfigurable system efficiently both in time and space, and have higher performance compared with other algorithms.

Sign in / Sign up

Export Citation Format

Share Document