Enables 3rd party network adapters and other devices to directly read and write CUDA host and device memory on NVIDIA Tesla™ and Quadro™ products. GPUDirect technology also includes direct transfers between GPUs.
The open source compiler infrastructure on which NVIDIA's CUDA Compiler (NVCC) is based on. Developers can create or extend programming languages with support for GPU acceleration using the CUDA Compiler SDK.
MPI is the industry standard API to enable applications threads to communicate across compute nodes. The latest implementations of this technology now support GPU accelerated nodes.