Fast Fourier Transform Implementation by Runtime code generation through Jumbo
The Fast Fourier Transform (FFT) is an algorithm that transforms a function from the time domain to the frequency domain. For many years it has been the subject of optimization efforts. We will attempt to optimize this algorithm through run-time code generation using the Jumbo compiler. Run-time code generation (programs creating programs at execution time) is a technique for exploiting dynamic information, taking advantage of data not yet available until runtime. Jumbo is a "two level" Java compiler which can create run-time code generators for Java. A previous version of the FFT algorithm was previously developed. This project further optimizes that code through data structure changes.
School:
University of Illinois at Urbana-Champaign
Department:
Computer Science
Research Advisor:
Sam Kamin
Department of Research Advisor:
Computer Science
Year of Publication:
2005