Speaker Deck

Embedded Languages for High-Performance Computing in Haskell

by Manuel Chakravarty

Published August 6, 2013 in Research

Embedded languages are a convenient and expressive method to capture patterns of high-performance code in functional languages. These patterns can be turned into efficient low-level code by template instantiation of code skeletons, where code fusion combines individual skeleton instances to minimise the abstraction penalty.

In this talk, I will illustrate these concepts as used in Accelerate, an embedded language for general-purpose GPU computing in Haskell that delivers competitive performance with a fraction of the effort required in low-level GPGPU frameworks, such as CUDA or OpenCL.