S. Ghiasi, K. Nguyen, E. Bozorgzadeh, and M. Sarrafzadeh (USA)
Delay Budgeting, Accuracy Improvement, Timing Constraint, Ap proximation, Computation Management
This paper presents the idea of managing the comprising computations of an application performed by an embedded networked system. An efficient algorithm for exploiting the timing slack of building blocks of the applica tion is proposed. The slack of blocks can be utilized by replacing them with slower but cheaper, i.e. better, modules and by assigning the computations to the proper resources. Thus, our approach manages the comprising computa tions and system resources and can indirectly assist the realtime scheduling of computations on system resources. This is performed without compromising the timing constraints of the application and can lead to significant improve ments in power dissipation, computation accuracy or other metrics of the application domain. Our algorithm is well-suited for arbitrary tree computa tions. Moreover, it delivers solutions that are desirably close to the optimal solution. Experimental results for a number of object tracking applications implemented in an networked system with embedded computation resources, exhibit a significant amount of slack utilization.
Important Links:
Go Back
