Generating FPGA-Accelerated DFT Libraries
Authors:
- Peter A. MilderPMPeter A. Milder
- Jeremy R. JohnsonJJJeremy R. Johnson
- 1 April 2007
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 173-184
Abstract
We present a domain-specific approach to generate high-performance hardware-software partitioned implementations of the discrete Fourier transform (DFT) in fixed point precision. The partitioning strategy is a heuristic based on the DFT's divide-and-conquer algorithmic structure and fine tuned by the feedback-driven exploration of candidate designs. We have integrated this approach in the Spiral linear-transform code-generation framework to support push-button automatic implementation. We present evaluations of hardware-software DFT implementations running on the embedded PowerPC processor and the reconfigurable fabric of the Xilinx Virtex-II Pro FPGA. In our experiments, the 1D and 2D DFT's FPGA-accelerated libraries exhibit between 2 and 7.5 times higher performance (operations per second) and up to 2.5 times better energy efficiency (operations per Joule) than the software-only version.Keywords
This publication has 14 references indexed in Scilit:
- Fast and accurate resource estimation of automatically generated custom DFT IP cores2006
- Formal loop merging for signal transforms2005
- Image Processing Application Development: From Rapid Prototyping to SW/HW Co-simulation and Automated Code GenerationLecture Notes in Computer Science, 2005
- A Scalable Embedded JPEG2000 ArchitectureLecture Notes in Computer Science, 2005
- Rewriting2001
- Exploiting FPGA-features during the emulation of a fast reactive embedded system1999
- SpecSyn: an environment supporting the specify-explore-refine paradigm for hardware/software system designIEEE Transactions on Very Large Scale Integration (VLSI) Systems, 1998
- Computational Frameworks for the Fast Fourier Transform1992
- A simple fixed-point error bound for the fast Fourier transformIEEE Transactions on Acoustics, Speech, and Signal Processing, 1979
- A fixed-point fast Fourier transform error analysisIEEE Transactions on Audio and Electroacoustics, 1969