A Hardware Task Scheduler for Embedded Video Processing
HiPEAC '09 Proceedings of the 4th International Conference on High Performance Embedded Architectures and Compilers
Proceedings of the Conference on Design, Automation and Test in Europe
| Hi-index | 0.00 |
The physical layer of most wireless protocols is traditionally implemented in custom hardware to satisfy the heavy computational requirements while keeping power consumption to a minimum. These implementations are time consuming to design and difficult to verify. A programmable hardware platform capable of supporting software implementations of the physical layer, or Software Defined Radio (SDR), has a number of advantages. These include support for multiple protocols, faster time-to-market, higher chip volumes, and support for late implementation changes. The challenge is to achieve this under the power budget of a mobile device. Wireless communications belong to an emerging class of applications with the processing requirements of a supercomputer but the power constraints of a mobile device — mobile supercomputing. This thesis presents a set of design proposals for building a programmable wireless communication solution. In order to design a solution that can meet the lofty requirements of SDR, this thesis takes an application-centric design approach — evaluate and optimize all aspects of the design based on the characteristics of wireless communication protocols. This includes a DSP processor architecture optimized for wireless baseband processing, wireless algorithm optimizations, and language and compilation tool support for the algorithm software and the processor hardware. This thesis first analyzes the software characteristics of SDR. Based on the analysis, this thesis proposes the Signal-Processing On-Demand Architecture (SODA), a fully programmable multi-core architecture that can support the computation requirements of third generation wireless protocols, while operating within the power budget of a mobile device. This thesis then presents wireless algorithm implementations and optimizations for the SODA processor architecture. A signal processing language extension (SPEX) is proposed to help the software development efforts of wireless communication protocols on SODA-like multi-core architecture. And finally, the SPIR compiler is proposed to automatically map SPEX code onto the multi-core processor hardware.