[ANNOUNCE] Parallel Haskell project underway
Parallel Haskell project underway GHC HQ and Well-Typed are pleased to report that work has started on the MSR-funded project to push the real-world use of parallel Haskell. We will be working with four industrial partners over the next two years, with the aim of demonstrating that parallel Haskell can be employed successfully in industrial projects. The participating organizations are: * Dragonfly (New Zealand) * IIJ Innovation Institute Inc. (Japan) * Los Alamos National Laboratory (USA) * Willow Garage Inc. (USA) Each group is working on their own project, applying parallel Haskell and their domain-specific expertise. In addition to providing advice on Haskell tools and techniques, we will work with these partners to identify and resolve any issues that are hindering progress. We are prepared to handle issues covering anything from the compiler and runtime system, through to platform, tool and library problems. All the participants are working on complex, real-world problems. Three projects involve scientific problems, and the fourth involves network servers. Three of the projects are targeting single-node SMP systems, while the fourth is targeting clusters. In two cases, Haskell will be directly pitted against existing code written in C or C++. Project progress reports will be posted to the Well-Typed blog [1] and to the new Parallel Haskell mailing list [2]. [1]: http://www.well-typed.com/blog/ [2]: http://groups.google.com/group/parallel-haskell == Dragonfly == http://www.dragonfly.co.nz/ Participants: Finlay Thompson, Edward Abraham Cloudy Bayes: Hierarchical Bayesian modeling in Haskell The Cloudy Bayes project aims to develop a fast Bayesian model fitter that takes advantage of modern multiprocessor machines. It will support model descriptions in the BUGS model description language (WinBUGS, OpenBUGS, and JAGS). It will be implemented as an embedded domain specific language (EDSL) within Haskell. A wide range of model hierarchical Bayesian model structures will be possible, including many of the models used in medical, ecological, and biological sciences. Cloudy Bayes will provide an easy to use interface for describing models, running Monte Carlo Markov chain (MCMC) fitters, diagnosing performance and convergence criteria as it runs, and collecting output for post-processing. Haskell's strong type system will be used to ensure that model descriptions make sense, providing a fast, safe development cycle. == IIJ Innovation Institute Inc. == http://www.iij-ii.co.jp/en/ Participants: Kazu Yamamoto Haskell is suitable for many kinds of domain, and GHC's support for lightweight threads makes it attractive for concurrency applications. An exception has been network server programming because GHC 6.12 and earlier have an IO manager that is limited to 1024 network sockets. The upcoming GHC 7 has a new IO manager implementation that gets rid of this limitation. This project will implement several network servers to demonstrate that Haskell is suitable for network servers that handle a massive number of concurrent connections. == Los Alamos National Laboratory == http://www.lanl.gov/ Participants: Michael Buksas, Timothy M. Kelley This project will use parallel Haskell to implement high-performance Monte Carlo algorithms, a class of algorithms which use randomness to sample large or otherwise intractable solution spaces. The initial goal is a particle-based MC algorithm suitable for modeling the flow of radiation, with application to problems in astrophysics. From this, the project is expected to move to identification of suitable abstractions for expressing a wider variety of Monte Carlo algorithms, and using models for different physical phenomena. == Willow Garage == http://www.willowgarage.com/ Participants: Ryan Grant Distributed Rigid Body Dynamics in ROS Willow Garage seeks a high-level representation for a distributed rigid body dynamics simulation, capable of excellent parallel speedup on current and foreseeable hardware, yet linking to existing optimized libraries for low-level message passing and matrix math. This project will drive API, performance, and profiling tool requirements for Haskell's interface to the Message Passing Interface (MPI) specification, an industry-standard in High Performance Computing (HPC), as used on clusters of many nodes. Competing internal initiatives use C++/MPI and CUDA directly. Willow Garage aims to lay the groundwork for personal robotics applications in everyday life. ROS (Robot Operating System - ROS.org) is an open source, meta-operating system for your robot. -- Dmitry Astapov Well-Typed LLP, http://www.well-typed.com/
On 15/11/10 15:23, Dmitry Astapov wrote:
== Dragonfly ==
Participants: Finlay Thompson, Edward Abraham
Cloudy Bayes: Hierarchical Bayesian modeling in Haskell
The Cloudy Bayes project aims to develop a fast Bayesian model fitter that takes advantage of modern multiprocessor machines. It will support model descriptions in the BUGS model description language (WinBUGS, OpenBUGS, and JAGS). It will be implemented as an embedded domain specific language (EDSL) within Haskell. A wide range of model hierarchical Bayesian model structures will be possible, including many of the models used in medical, ecological, and biological sciences.
This is slightly OT, but having used WinBUGS a little, I am very pleased to see this listed as a project. WinBUGS is a commendable piece of work, but is really showing its age these days. Between its really horrible error messages if anything goes wrong, awkward scripting/job-running interface (slightly eased by R2WinBUGS), and of course lack of support for multicore, I think this is a case where parallel Haskell could provide a lot of benefits over the existing software. Also, can you tell us which (if any) of these projects will be released as open-source? Thanks, Neil.
On Mon, Nov 15, 2010 at 5:44 PM, Neil Brown
On 15/11/10 15:23, Dmitry Astapov wrote:
[snip]
This is slightly OT, but having used WinBUGS a little, I am very pleased to see this listed as a project. WinBUGS is a commendable piece of work, but is really showing its age these days. Between its really horrible error messages if anything goes wrong, awkward scripting/job-running interface (slightly eased by R2WinBUGS), and of course lack of support for multicore, I think this is a case where parallel Haskell could provide a lot of benefits over the existing software.
This is very interesting, thank you for your input!
Also, can you tell us which (if any) of these projects will be released as open-source?
This is a multi-faceted issue. First of all, all work done in the scope of this projects on the existing tools (like GHC, Cabal, various other packages, ...) would be released under the respective licenses. If, during the course of the project, we would identify the need to develop libraries and tools that would be of general utility, we would strive to release them under open-source licenses. What would happen with anything developed for the specific need of project participants would be up to them, but then such sources would probably be of little interest to public anyway. -- Dmitry Astapov Well-Typed LLP, http://www.well-typed.com/
participants (2)
-
Dmitry Astapov -
Neil Brown