An Embedded Language for Accelerated Array Computations
Data.Array.Accelerate defines an embedded language of array computations for high-performance computing in Haskell. Computations on multi-dimensional, regular arrays are expressed in the form of parameterised collective operations (such as maps, reductions, and permutations). These computations are online-compiled and executed on a range of architectures.
For more details, see our recent paper Accelerating Haskell Array Codes with Multicore GPUs. There are also some slightly outdated slides and a video of a talk that I gave at the Haskell Implementors Workshop 2009 (in Edinburgh): Haskell Arrays, Accelerated (Using GPUs).
A simple example
As a simple example, consider the computation of a dot product of two vectors of single-precision floating-point numbers:
dotp :: Acc (Vector Float) -> Acc (Vector Float) -> Acc (Scalar Float) dotp xs ys = fold (+) 0 (zipWith (*) xs ys)
Except for the type, this code is almost the same as the corresponding Haskell code on lists of floats. The types indicate that the computation may be online-compiled for performance — for example, using
Data.Array.Accelerate.CUDA.run it may be on-the-fly off-loaded to a GPU.
Package accelerate is available from
- Hackage: accelerate — install with
cabal install accelerate
- GitHub: AccelerateHS/accelerate - get the source with
git clone https://github.com/AccelerateHS/accelerate.git
The following supported addons are available as separate packages on Hackage and included as submodules in the GitHub repository:
accelerate-cudaBackend targeting CUDA-enabled NVIDA GPUs
accelerate-examplesComputational kernels and applications showcasing the use of Accelerate as well as a regression test suite (supporting function and performance testing)
accelerate-ioFast conversion between Accelerate arrays and other array formats (including Repa arrays)
accelerate-backend-kitSimplified internal AST to get going on writing backends
accelerate-buildbotBuild bot for automatic performance & regression testing
The following additional components are experimental and incomplete:
accelerate-openclBackend targeting GPUs via the OpenCL standard
accelerate-repaBackend targeting multicore CPUs via the Repa parallel array library
- Glasgow Haskell Compiler (GHC), 7.0.3 or later
- Haskell libraries as specified in
- For the CUDA backend, CUDA version 3.0 or later
Examples and documentation
The GitHub repository contains a submodule accelerate-examples, which provides a range of computational kernels and a few complete applications. To install these from Hackage, issue
cabal install accelerate-examples.
- Haddock documentation is included in the package and linked from the Hackage page.
- Online documentation is on the GitHub wiki.
- The idea behind the HOAS (higher-order abstract syntax) to de-Bruijn conversion used in the library is described separately.
The maintainer of this package is Manuel M T Chakravarty email@example.com (aka TacticalGrace on #haskell and related channels).
Both user and developer questions and discussions are welcome at
firstname.lastname@example.org. Sorry, this mailing list is currently unavailable.
Bug reports and issues tracking are on the GitHub project page.
Here is a list of features that are currently missing:
- The CUDA backend does not support arrays of type Char and Bool at the moment.
- Preliminary API (parts of the API may still change in subsequent releases)