In principle Haskell is a great language for parallel programming (few side effects, implicit parallelism, yada yada), but how does this work out in practice? This tutorial will put you in a position where you can try it for yourself. There will be three parts (about an hour each):
We will mostly concentrate on parallel programming (where the goal is to improve performance) rather than concurrent programming (where the goal is modular program design).
Take home: A good understanding of the basics of two parallel-programming paradigms in Haskell.
Generally we will assume that you are reasonably proficient in Haskell. However, while non-Haskellers may not get all the details, you should still have a good time. The "whirlwind tour" part in particular requires minimal Haskell expertise.
From the command line:
apt-get install Haskell-platform
apt-get install threadscope
cabal install monad-par
This gets you GHC 6.12.3 at the moment, which is enough to run the examples and do the exercises in the tutorial.
Alternatively, you can get a more up to date Haskell Platform here and installing it yourself.
You may also want to get the newer version of ThreadScope (0.2). For instructions go to this page.
Install the Haskell Platform.
Install the monad-par package (using the command line):
cabal install monad-par
Also, an installation of ThreadScope is optional but recommended.
Install the Haskell Platform.
Install the monad-par package (using the command line):
cabal install monad-par
ThreadScope is optional but recommended, although it depends on the GTK library which is reportedly hard to install on Mac OS X. Apparently installing it via HomeBrew is the method most likely to work. Some tips on installing ThreadScope are on the ThreadScope page.
is a Principal Researcher at Microsoft's research lab in Cambridge England. He became addicted to functional programming in 1979, and has been working on the design and implementation of functional languages ever since. He is one of the original designers of Haskell and with Simon Marlow is responsible for the Glasgow Haskell Compiler (GHC).
is a Senior Research Software Engineer at Microsoft's Cambridge research lab. He has been working on Haskell-related research projects since the 90's, and together with Simon Peyton Jones is responsible for the Glasgow Haskell Compiler (GHC). Simon in particular developed GHC's multicore runtime system, and over the past few years has been working to improve GHC's support for parallel and concurrent programming, both at the runtime and the language level.
is an Associate Professor at the University of New South Wales, Sydney. His main research interests are in functional and parallel programming. He contributed to Haskell's foreign function interface, the theory and implementation of type families, and the design and realisation of Data Parallel Haskell, an implementation of nested data parallelism in the Glasgow Haskell Compiler.