DESIGN AND IMPLEMENTATION OF MIMO OFDM IEEE 802.11N RECEIVER BLOCKS ON HETEROGENEOUS MULTICORE ARCHITECTURE

In this thesis, the performance of a heterogeneous multicore platform in terms of technical capability is evaluated. Therefore, the choice of architecture in general can be based on a set of diverse applications. Selected applications can be parallel or serial in nature. Applications evaluation are often based on various performance metrics including the resource utilization and execution time. The wireless communication systems are expanded to accelerate their functions execution in both software and hardware. The embedded systems which involve several types of communication systems perform a large number of computations which require short execution time and minimized power consumption. Also, there is a growing demand for application-specific accelerators aiding general purpose. One feasible way is to use heterogeneous multi-core platforms. Furthermore, many application-specific accelerators are loosely connected with each other. In this study, the implementation of Multiple-Input Multiple-Output (MIMO) Orthogonal Frequency Division Multiplexing (OFDM) receiver is evaluated by applying a Heterogeneous Multicore Architecture (HMA). The MIMO OFDM receiver is composed of computationally intensive and general-purpose processing tasks and can serve maximum coverage for evaluation of the HMA. The receiver blocks are designed by crafting template-based Coarse-grained Reconfigurable Array (CGRA) devices. In this case study, four streams (antennas) are proposed in order to process the data over CGRAs simultaneously. HMA nodes will be reconfigured at run-time in different blocks of the receiver. In this experimental work, according to the performance of each CGRA, the collective performance of the entire platform as well as NoC traffic is recorded considering the number of clock cycles and also several high-level performance criteria. The implementation of OFDM receiver scaled CGRAs to various dimensions. The data can also be exchanged between diverse nodes on the NoC structure by utilizing direct memory access (DMA) devices independently.