Section: Research Program

Introduction

The main research topic of the DaRT team-project concerns the hardware/software codesign of embedded systems with high performance processing units like DSP or SIMD processors. A special focus is put on multi processor architectures on a single chip (System-on-Chip). The contribution of DaRT is organized around the following items:

Co-modeling for High Performance SoC design:

We define our own metamodels to specify application, architecture, and (software hardware) association. These metamodels present new characteristics as high level data parallel constructions, iterative dependency expression, data flow and control flow mixing, hierarchical and repetitive application and architecture models. All these metamodels are implemented with respect to the MARTE standard profile of the OMG group, which is dedicated to the modeling of embedded and real-time systems.

Model-based optimization and compilation techniques:

We develop automatic transformations of data parallel constructions. They are used to map and to schedule an application on a particular architecture. This architecture is by nature heterogeneous and appropriate techniques used in the high performance community can be adapted. We developed new heuristics to minimize the power consumption. This new objective implies to specify multi criteria optimization techniques to achieve the mapping and the scheduling.

SoC simulation, verification and synthesis:

We develop a SystemC based simulation environment at different abstraction levels for accurate performance estimation and for fast simulation. To address an architecture and the applications mapped on it, we simulate in SystemC at different abstraction levels the result of the SoC design. This simulation allows us to verify the adequacy of the mapping and the schedule, e.g., communication delay, load balancing, memory allocation. We also support IP (Intellectual Property) integration with different levels of specification. On the other hand, we use formal verification techniques in order to ensure the correctness of designed systems by particularly considering the synchronous approach. Finally, we transform MARTE models of data intensive algorithms in VHDL, in order to synthesize a hardware implementation.