GOM-OpenCL aims to implement a fully parallelized version of the P3 algorithm by Brian W. Goldman and William F. Punch using the raw power of modern GPUs.
The implementation relies on the OpenCL bindings provided by rainliu:
go get -v -tags="cl12" github.com/rainliu/opencl/cl
To build/install the executable from the repository execute the following command:
go get -v -tags="cl12" github.com/Morenim/gom-opencl
Only the latest OpenCL 1.2 specification is officially supported.
The executable was tested on hosts with the following specifications:
CPU: Core i7-2600k clocked at ~3.4Ghz
GPU: Nvidia GTX 560 Ti clocked at 882 Mhz with 384 CUDA cores and 1024 MB GDDR5 memory.
RAM: 16 GB DDR3 clocked at 1333 Mhz