HaskSkel: Algorithmic Skeletons in Haskell

Kevin Hammond, AJR Portillo, P Koopman, C Clack

Research output: Contribution to conferencePaper

9 Citations (Scopus)


This paper introduces HaskSkel: an implementation of algorithmic skeletons for the non-strict functional language Haskell. The implementation is based on the evaluation strategy mechanism for Glasgow Parallel Haskell (GpH), and has been tested in both simulated and real parallel settings. Compared with most skeleton libraries, HaskSkel has a number of novel and interesting features: (strict) skeletons have been embedded in a non-strict language; the cost models and implementation modules are written entirely using standard Haskell (and may therefore be modified or extended by the user rather than being a fixed part of the compiler); it is possible to nest skeletons (though we have not yet defined nested cost models to support this); and the library is not restricted to a single bulk data type. Finally, in contrast with most other skeletal approaches, it is possible to freely mix HaskSkel skeletons with other parallel constructs that have been introduced using standard evaluation strategies.

Original languageEnglish
Publication statusPublished - 2000


Dive into the research topics of 'HaskSkel: Algorithmic Skeletons in Haskell'. Together they form a unique fingerprint.

Cite this