Update: August 30th, 2017
The exact optimized parallelization for GPU, Multicore , DSP is not updated on the public repository. As i am writing a paper to publish it. As soon as I submit my papers, i will publish the optimized OpenSource Code. :)
The OpenACC Sparse Fast Fourier Transform Library
Version 1.0 April 2017
This is the Sparse Fast Fourier Transform Library, a library to compute Discrete Fourier Transforms of signals with a sparse frequency domain. Now with OpenACC API to port the code to GPUs and increase the performance.
- AUTHORS
The original SFFT sourcecode was developed by Haitham Hassanieh, Piotr Indyk, Dina Katabi, and Eric Price at the Computer Science and Artifical Intelligence Lab at MIT. The original sourcecode and contact information can be found at their website [1].
Performance optimizations were developed by Jörn Schumacher joerns@student.ethz.ch as part of his master thesis project [2] at the Computer Science Department of ETH Zurich in 2013, under the supervision of Prof. Markus Püschel [3].
[1] Sparse Fast Fourier Transform, http://groups.csail.mit.edu/netmit/sFFT/. [2] Jörn Schumacher, "High-Performance Sparse Fast Fourier Transform", Master thesis, Computer Science, ETH Zurich, Switzerland, 2013. [3] Homepage of Markus Püschel, http://www.inf.ethz.ch/personal/markusp/.
Parallelizations of the code using OpenACC API's is done by Arnov Sinha arnov@udel.edu as a part of his Master thesis project at the Computer and Information Science Department of University of Delaware, USA in 2017, under the supervision of Prof. Sunita Chandrasekaran.
- CONTACT INFORMATION
If you are interested in the theory behind the Sparse Fast Fourier Transform, contact the inventors of the SFFT, Haitham Hassanieh, Piotr Indyk, Dina Katabi, and Eric Price, at their website http://groups.csail.mit.edu/netmit/sFFT/.
If you are interested in performance optimizations that were applied, contact Jörn Schumacher at joerns@student.ethz.ch.
If you are interested in the parallelization methods and performance gains using HPC, contact Arnov Sinha arnov@udel.edu.
- DISCLAIMER
The current SFFT implementation is in an experimental state. It is NOT intended to be used as a drop-in replacement for the FFT library of your choice. Be prepared to find bugs. There is absolutely NO WARRANTY for the correct functioning of this software.
- License
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- Installation
For installation instructions and a full manual, build the documentation using "make html" in the doc/ directory.
Usually, a "./configure && make && make install" is enough to install the library. This will build a static and a shared variant of the SFFT library. Note that by default you need the Intel IPP library; to disable this, pass '--without-ipp' to the configure step.