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 There are also some slightly outdated slides and a video of a talk that I gave at the Haskell Implementors Workshop 2009 (in Edinburgh): Haskell Arrays, Accelerated (Using GPUs).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.

Availability

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

Additional components

The following supported addons are available as separate packages on Hackage and included as submodules in the GitHub repository:

• accelerate-cuda Backend targeting CUDA-enabled NVIDA GPUs
• accelerate-examples Computational kernels and applications showcasing the use of Accelerate as well as a regression test suite (supporting function and performance testing)
• accelerate-io Fast conversion between Accelerate arrays and other array formats (including Repa arrays)
• accelerate-backend-kit Simplified internal AST to get going on writing backends
• accelerate-buildbot Build bot for automatic performance & regression testing

The following additional components are experimental and incomplete:

• accelerate-opencl Backend targeting GPUs via the OpenCL standard
• accelerate-repa Backend targeting multicore CPUs via the Repa parallel array library

Requirements

• Glasgow Haskell Compiler (GHC), 7.0.3 or later
• Haskell libraries as specified in accelerate.cabal
• 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.

Contacts

The maintainer of this package is Manuel M T Chakravarty chak@cse.unsw.edu.au (aka TacticalGrace on #haskell and related channels).

Both user and developer questions and discussions are welcome at accelerate@projects.haskell.org. Sorry, this mailing list is currently unavailable.

Bug reports and issues tracking are on the GitHub project page.

What's missing?

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)