Stream Processing for Embedded Domain Specific Languages
Conference contribution, 2014

We present a library for expressing digital signal processing (DSP) algorithms using a deeply embedded domain-specific language (EDSL) in Haskell. The library supports definitions in functional programming style, reducing the gap between the mathematical description of streaming algorithms and their implementation. The deep embedding makes it possible to generate efficient C code without any overhead associated with the high-level programming model. The signal processing library is intended to be an extension of the Feldspar EDSL which, until now, has had a rather low-level interface for dealing with synchronous streams. However, the presented library is independent of the underlying expression language, and can be used to extend any pure EDSL for which a C code generator exists with efficient stream processing capabilities. The library is evaluated using example implementations of common DSP algorithms and the generated code is compared to its handwritten counterpart.


Digital Signal Processing



Markus Aronsson

Chalmers, Computer Science and Engineering (Chalmers), Software Technology (Chalmers)

Emil Axelsson

Chalmers, Computer Science and Engineering (Chalmers), Software Technology (Chalmers)

Mary Sheeran

Chalmers, Computer Science and Engineering (Chalmers), Software Technology (Chalmers)

26th Symposium on Implementation and Application of Functional Languages, IFL 2014, Northeastern University Boston, United States, 1 October 2014 through 3 October 2014


Areas of Advance

Information and Communication Technology

Subject Categories

Computer and Information Science





More information