Title: Generics in Data Parallel Haskell

Roman Leshchinskiy, University of New South Wales

Abstract:

The Data Parallel Haskell project is an ongoing effort to extend the
Glasgow Haskell Compiler with support for nested data parallelism, a
high-level, efficient approach to implementing parallel applications.
In its implementation, we make extensive use of generic programming
which allows us to transfer large amounts of functionality from the
compiler to the library. In this talk, we describe our approach and
explain why it is crucial for seamlessly integrating DPH with the rest
of GHC.