Co-synthesis of Heterogeneous Multi-Task Embedded Systems
with Real Time Constraints
In this project, we investigate various co-synthesis
algorithms that synthesize heterogeneous multiprocessor embedded systems with
real-time constraints. The system partitioning is performed using Genetic
algorithm. The proposed co-synthesis method must explore target architecture
solutions from a number of available processing elements. Each of these
solutions is an optimized architecture for the system. The co-synthesis
algorithm should also be capable of producing feasible solutions for
large-scale systems. Scheduling of the processes is performed by a new
non-preemptive technique. The scheduling technique can deal with system level
as well as process level real-time deadlines.
The co-synthesis method should enable the user to produce target system architecture requiring fewer processing element .It also tries to achieve better processing element utilization as compared to other methods presented in the literature.