GpH -- Glasgow Parallel Haskell

Activity Report November 2004

Report by: Phil Trinder

The Team

Heriot-Watt University, Scotland: Phil Trinder, Abyd Al Zain, Andre Rauber du Bois

St Andrews University, Scotland: Kevin Hammond, Leonid Timochouk, Yang Yang

Phillips Universität Marburg, Germany: Jost Berthold

Brooklyn College, USA: Murray Gross

Status

A complete, GHC-based implementation of the parallel Haskell extension GpH and of evaluation strategies is available.

System Evaluation and Enhancement

The first 3 items are linked by a British Council/DAAD collaborative project between Heriot-Watt University, St Andrews University, and Phillips Universität Marburg.

We are adapting GpH to run on computational GRIDs. The current implementation performs well on single clusters, and multiple clusters with a low-latency interconnect. A distribution is available on request from (ceeatia@macs.hw.ac.uk).
We are designing a generic parallel runtime environment encompassing both the Eden and GpH runtime environments
In separate work GpH is being used as a vehicle for investigating scheduling on the GRID.
We are teaching parallelism to undergraduates using GpH at Heriot-Watt and Phillips Universitat Marburg.

GpH Applications

GpH is being used to parallelise the GAP mathematical library in an EPSRC project (GR/R91298).

Implementations

The GUM implementation of GpH is available in two development branches, and work on a port of GUM to the latest GHC 6.xx branch has been started over summer.

The stable branch (GUM-4.06, based on GHC-4.06) is available for RedHat-based Linux machines: binary snapshot (see installation instructions). The stable branch is available from the GHC CVS repository via tag gum-4-06.
The unstable branch (GUM-5.02, based on GHC-5.02) is working and has been used on a Beowulf cluster. It is available on request as a source bundle.

Our main hardware platform are Intel-based Beowulf clusters. Work on ports to other architectures is also moving on (and available on request). Specifically a port to a Mosix cluster has been built in the Metis project at Brooklyn College, with a first version available on request from Murray Gross.