A Layered Modeling and Simulation Approach to investigate Resource-aware Computing in MPSoCs

TL;DR

This paper introduces a modular modeling and simulation approach for resource-aware computing in MPSoCs, enabling early-stage design exploration to improve system performance and resource utilization.

Contribution

It presents a layered methodology that separates resource-awareness modeling from system modeling, facilitating rapid exploration of design options in MPSoCs.

Findings

Effective resource-aware execution exploration demonstrated

Modular approach enables rapid design space exploration

Reusability of legacy models enhances efficiency

Abstract

Increasing complexity of modern multi-processor system on chip (MPSoC) and the decreasing feature size have introduced new challenges. System designers have to consider now aspects which were not part of the design process in past times. Resource-aware Computing is one of such emerging design concerns which can help to improve performance, dependability and resource utilization of overall system. Resource-aware execution takes into account the resource status when executing tasks on MPSoCs. Exploration of resource-aware computing at early design stages of complex systems is mandatory and appropriate methodologies to do this in an efficient manner are thus required. In this paper, we present a modular approach which provides modeling and simulation support for investigation of resource-aware execution in MPSoCs. The proposed methodology enables rapid exploration of the design space by modeling and simulating the resource-awareness in a separate layer while widely reusing the legacy system model in the other layer. Our experiments illustrate the benefits of our approach for the exploration of resource-aware execution on MPSoCs.