Synthesis of Application Specific Multiprocessor Architectures for Process Networks

Authors:
Basant Kumar Dwivedi;Anshul Kumar;M. Balakrishnan
Affiliations:
-;-;-
Venue:
VLSID '04 Proceedings of the 17th International Conference on VLSI Design
Year:
2004

Citing 0
Cited 2

Streamroller:: automatic synthesis of prescribed throughput accelerator pipelines

CODES+ISSS '06 Proceedings of the 4th international conference on Hardware/software codesign and system synthesis
Optimal synthesis of latency and throughput constrained pipelined MPSoCs targeting streaming applications

CODES/ISSS '10 Proceedings of the eighth IEEE/ACM/IFIP international conference on Hardware/software codesign and system synthesis

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we address the problem of synthesis of applicationspecific multiprocessor SoC architectures for processnetworks of streaming applications. An application ismodeled as Kahn Process Network (KPN) which makes theparallelism present in the application explicit. The synthesisprocess involves selection of computation modules, memorymodules, communication architecture and mapping of processesof KPN on compute units and channels on memorymodules. Our solution minimizes hardware cost while takinginto account the performance constraints. One of thesalient features of our work is that it takes into account theadditional overheads because of data communication conflicts.Our method uses average processing requirements ofKPN to handle data dependent behavior of processes andcycles within the KPN. In contrast to others, we do not performstatic scheduling, only mapping and synthesis is done.