Glasgow Parallel Haskell

Around 1990 the par and pseq primitives were used in a parallel extension of Lazy ML (LML) on the novel GRIP parallel graph reduction machine (Hammond and Peyton-Jones 1990).
From 1995 onwards the GranSim Profiler was developed for GpH programs. GranSim was extended as GranCC for parallel cost centre profiling, and GranSP for profiling strategies.
In 1996 evaluation strategies that abstract over the par and pseq primitives were developed for the GUM distributed-memory parallel Haskell implementation (Trinder et al. 1998).
Around 2000 evaluation strategies are used alongside processes in the Eden parallel Haskell variant (Klusik et al. 2000).
In 2000 Glasgow Distributed Haskell (GdH) is designed as a superset of GpH supporting distributed stateful threads.
In 2001 systematic generic data clustering was developed to improve locality and granularity (Loidl et al. 2001).
In 2008 evaluation strategies provide effective parallelism for the GHC shared memory parallel Haskell implementation (Marlow et al. 2009).
In 2010 a new version of evaluation strategies was developed that are more generic and more efficient and that can directly express embedded strategies (Marlow et al. 2010).

The History of GpH