Matching your domain mastery to our GPU expertise.
Microsoft’s C++ AMP (Accelerated Massive Parallelism) accelerates the performance of your C++ applications by utilizing the data-parallel hardware found in modern GPUs. Since C++ AMP is currently implemented on top of the DirectX 11 multimedia API, this technology can be utilized on any graphics accelerator that supports the DirectX 11 standard. Currently this includes GPUs from NVIDIA and AMD as well as support for CPU fallback if no accelerator is present.
By using modern C++ programming techniques such as lambda expressions and template methods, developers can quickly and cleanly develop applications without having to deal with the low level details of GPU programming such as explicit memory management and kernel configuration. This allows for rapid development in which GPU acceleration can be achieved with minimal development time.
The engineers at EM Photonics have been creating C++ AMP applications since the technology was unveiled by Microsoft. Our experience has informed our development of best practices for creating C++ AMP applications from scratch and porting existing applications. Through the use of features such as tiled operations and asynchronous transfers, our engineers can help you obtain optimal performance from the technology.
For more information or to further discuss your needs, please contact us directly.