GPGPU stands for General-purpose computing on graphics processing units.
In Linux, there are currently two major GPGPU frameworks: OpenCL and CUDA
OpenCL (Open Computing Language) is an open, royalty-free parallel programming framework developed by the Khronos Group, a non-profit consortium.
Distribution of the OpenCL framework generally constists of:
- Library providing OpenCL API, known as libCL or libOpenCL (libOpenCL.so in linux)
- OpenCL implementation(s), which contain:
- Device drivers
- OpenCL/C code compiler
- SDK *
- Header files *
* only needed for development
There are several choices for the libCL. In general case, installing Template:Package Official from [extra] should do :
# pacman -S libcl
However, there are situations when another libCL distribution is more suitable. The following paragraph covers this more advanced topic.
The OpenCL ICD model
OpenCL offers the option to install multiple vendor-specific implementations on the same machine at the same time. In practice, this is implemented using the Installable Client Driver (ICD) model. The center point of this model is the libCL library which in fact imeplements ICD Loader. Through the ICD Loader, an OpenCL application is able to access all platforms and all devices present in the system.
Although itself vendor-agnostic, the ICD Loader still has to be provided by someone. In Archlinux, there are currently two options:
- extra/Template:Package Official by Nvidia. Provides OpenCL version 1.0 and is thus slightly outdated. It's behaviour with OpenCL 1.1 code has not been tested as of yet.
- unsupported/Template:Package AUR by AMD. Provides up to date version 1.1 of OpenCL. It is currently distributed by AMD under a restrictive license and therefore could not have been pushed into official repo.
(There is also Intel's libCL, this one is currently not provided in a seperate package though.)
For basic usage, extra/libcl is recommended as it's installation and updating is convenient. For advanced usage, libopencl is recommended. Both libcl and libopencl should still work with all the implementations.
To see which OpenCL imeplentations are currently active on your system, use the following command: Template:Cli
OpenCL implementation from AMD is known as AMD APP SDK, formerly also known as AMD Stream SDK or ATi Stream.
For Arch Linux, AMD APP SDK is currently available in AUR as Template:Package AUR. This package is installed as Template:Filename and apart from SDK files it also contains a profiler (Template:Filename) and a number of code samples (Template:Filename). It also provides the Template:Filename utility which lists OpenCL platforms and devices present in the system and displays detailed information about them.
As AMD APP SDK itself contains CPU OpenCL driver, no extra driver is needed to use execute OpenCL on CPU devices (regardless of it's vendor). GPU OpenCL drivers are provided by the Template:Package AUR package (an optional dependency), the open-source driver (Template:Package Official) does not support OpenCL.
Code is compiled using Template:Package Official (dependency).
The Intel implementation, named simply Intel OpenCL SDK, provides optimized OpenCL performance on Intel CPUs (mainly Core and Xeon) and CPUs only. There is no GPU support as Intel GPUs don't support OpenCL/GPGPU. Package is available in AUR: Template:Package AUR.
For development of OpenCL-capable applications, full installation of the OpenCL framework including implementation, drivers and compiler plus the Template:Package Official package is needed. Link your code against libOpenCL.
- C++: A binding by Khronos is part of the official specs. It is included in Template:Package Official
- C++/Qt: An experimental binding named QtOpenCL is in Qt Labs - see Blog entry for more information
- Python: There are two bindings with the same name: PyOpenCL. One is in [extra]: Template:Package Official, for the other one see sourceforge
- D: cl4d
- Haskell: The OpenCLRaw package is available in AUR: Template:Package AUR
- Java: JOCL (a part of JogAmp)
- Mono/.NET: Open Toolkit
CUDA (Compute Unified Device Architecture) is Nvidia's proprietary, closed-source parallel computing architecture and framework. It is made of several components:
- proprietary Nvidia kernel module
- CUDA "driver" and "runtime" libraries
- additional libraries: CUBLAS, CUFFT, CUSPARSE, etc.
- CUDA toolkit, including the Template:Filename compiler
- CUDA SDK, which contains many code samples and examples of CUDA and OpenCL programs
The kernel module and CUDA "driver" library are shipped in extra/Template:Package Official and extra/Template:Package Official. The "runtime" library and the rest of the CUDA toolkit are available in unsupported/Template:Package AUR. The SDK has been packaged too (Template:Package AUR), even if it is not required for developing in CUDA.
- Fortran: FORTRAN CUDA, PGI CUDA Fortran Compiler
- Python: In AUR: Template:Package AUR, also Kappa
- Perl: Kappa, CUDA-Minimal
- Haskell: The CUDA package is available in AUR: Template:Package AUR. There is also The accelerate package
- Java: jCUDA, JCuda
- Mono/.NET: CUDA.NET, CUDAfy.NET
- Mathematica: CUDAlink
- Ruby, Lua: Kappa