Benchmarking implementations of lazy functional languages
Pieter Hartel, Koen G. Langendoen · 1993
Five implementations of different lazy functional languages are compared using a common benchmark of a dozen medium size programs. The benchmarking procedure has been designed such that one set of programs can be translated automatically into different languages, thus allowing a comparison of the quality of compilers for different lazy functional languages. Aspects studied include compile time, execution time, ease of programming determined by the availability of certain key features, and the quality of the documentation. All compilers studied generate good quality code. The Nijmegen Clean system compiles faster than all the others. The FAST/FCG compiler from Southampton/Amsterdam properly supports arrays. The LML system from Chalmers is the most robust. The Haskell compilers from Chalmers and Glasgow provide the most comprehensive functionality. 1 Introduction To take stock of the current state of affairs, a number of compilers for lazy functional languages have been benchmarked. We...