NOTE: This package is being deprecated in favour of accelerate-llvm, which supports execution on multicore CPUs as well as CUDA-capable GPUs. See the accelerate-llvm package for details.

This package compiles Accelerate code down to NVIDIA's CUDA language for general-purpose GPU programming. For details on Accelerate, refer to the main repository. Please also file bug reports and feature requests with the issue tracker of the main repository.

To use this package, you need a CUDA-enabled NVIDIA GPU and NVIDIA's CUDA SDK version 3.* or later. You can find the SDK at the NVIDIA Developer Zone. We recommend to use hardware with compute capability 1.2 or greater — see the table on Wikipedia.