fft algorithm implementation with hadoop

I want to implement Fast Fourier Transform algorithm with Hadoop. I know recursive-fft algorithm but I need your guideline in order to impleme开发者_如何转开发nt it Map/Reduce approach. Any suggestions?

Thanks.

I have a tentative solution here:

http://blog.jierenchen.com/2010/08/fft-with-mapreduce.html

I haven't tried coding this up so I'm not 100% sure it works. Lemme know if I made any booboos.

For using MapReduce to compute large scale FFTs, there is a detailed discussion in [1]. The corresponding presentation slides are available at [2]. The source code of a Hadoop implementation is available at [3].

[1] Schönhage-Strassen Algorithm with MapReduce for Multiplying Terabit Integers (SNC2011)

[2] http://www.slideshare.net/hortonworks/large-scale-math-with-hadoop-mapreduce

[3] https://issues.apache.org/jira/browse/MAPREDUCE-2471