Evaluations of hardware platforms, methods and tools for low-volume software-defined signal processing applications
(2011)
Files
thesis_release_20110218_final.pdf
Open Access - Adobe PDF
- 2.39 MB
Details
- Authors
- Supervisors
- Legat, Jean-Didier
- Abstract
- (en) Embedded signal processing applications evolve towards advanced applications supporting many standards and providing advanced functionalities. In those applications, the control and software parts are increasing and their implementations shift progressively from full hardware implementations to partial or full software implementations. In this thesis, those software implementations are regrouped under the name of software-defined signal processing applications, or SDAs. Moreover, among embedded electronics, many applications have low volumes of production. For those applications, the non-recurring engineering costs (NREs) related to application development and platform realization are a major issue. Conventional hardware platforms for signal processing applications like heterogeneous platforms cannot meet all the requirements of SDAs and low-volume applications. This situation has motivated the development of new hardware platform models like multi-core processors, SIMD processors, coarse-grain reconfigurable platforms, fine-grain reconfigurable platforms like FPGAs, and homogeneous many-core platforms (HMCP). In this thesis, those five hardware platform models are evaluated and compared. Two platform models are further studied: FPGAs with dynamic partial reconfiguration and HMCPs.
- Affiliations
Citations
Rousseau, B. (2011). Evaluations of hardware platforms, methods and tools for low-volume software-defined signal processing applications. https://hdl.handle.net/2078.5/149421