Integration of dataflow optimization techniques into a software radio design framework

Authors:
George F. Zaki;William Plishker;Tim Oshea;Nick McCarthy;Charles Clancy;Eric Blossom;Shuvra S. Bhattacharyya
Affiliations:
Electrical and Computer Engineering Department, University of Maryland, College Park, Maryland;Electrical and Computer Engineering Department, University of Maryland, College Park, Maryland;Laboratory for Telecommunications Sciences, University of Maryland, College Park, Maryland;Laboratory for Telecommunications Sciences, University of Maryland, College Park, Maryland;Laboratory for Telecommunications Sciences, University of Maryland, College Park, Maryland;Blossom Research, LLC, Reno, NV;Electrical and Computer Engineering Department, University of Maryland, College Park, Maryland
Venue:
Asilomar'09 Proceedings of the 43rd Asilomar conference on Signals, systems and computers
Year:
2009

Citing 7
Cited 0

Static scheduling of synchronous data flow programs for digital signal processing

IEEE Transactions on Computers
GNU radio: tools for exploring the radio frequency spectrum

Linux Journal
Software synthesis from the dataflow interchange format

SCOPES '05 Proceedings of the 2005 workshop on Software and compilers for embedded systems
Hierarchical coarse-grained stream compilation for software defined radio

CASES '07 Proceedings of the 2007 international conference on Compilers, architecture, and synthesis for embedded systems
Functional DIF for Rapid Prototyping

RSP '08 Proceedings of the 2008 The 19th IEEE/IFIP International Symposium on Rapid System Prototyping
A generalized scheduling approach for dynamic dataflow applications

Proceedings of the Conference on Design, Automation and Test in Europe
Cycle-static dataflow

IEEE Transactions on Signal Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Application specific design frameworks, such as GNU Radio for software defined radio, facilitate fast design flows by leveraging the common application structures of particular domains and rich libraries of elements tailored to them. However, due to their focus on the application, implementations derived from these frameworks are often unable to take advantage of target platform features to achieve high performance. To apply more extensive optimizations, applications described in such frameworks can be refined to formal models, making them more amenable to analysis and optimization. In this work, we present a method for integrating GNU Radio with a dataflow design framework, the Dataflow Interchange Format (DIF). We describe a translation from GNU Radio applications into a formal dataflow model and the software infrastructure we have used to integrate the two software packages. This integration allows GNU Radio to employ a variety of dataflow schedulers to improve performance on existing applications. Furthermore, by applying formal models to the application and the target architecture, this integration should allow for target specific optimizations for additional performance and target flexibility.