Microsoft has revealed its plans to include possibilities for heterogeneous computing into next version of C++ compiler.
Today, in his speech at AMD Fusion Developer Summit, S. Somasegar, Microsoft Developer Division Senior Vice President, has introduced C++ Accelerated Massive Parallelism (C++ AMP) technology.
This technology is based on DirectX platform and will be fully integrated into Visual Studio vNext — next version of Visual Studio (most likely, a public release will be called Visual Studio 2012).
The main task for C++ AMP technology is to use GPU compute powers. Developers will get a quite powerful tool for parallel compute power allocation between CPU and GPU. Besides compiler, PPL and C++ Runtime execution libraries will get updates with support for GPGPU technology too.
As Microsoft notes, C++ AMP specification will be open specification, which will allow it to be used in other compilers. Apart from Windows version, this technology will be available in compilers of non-Windows platforms as well.
Daniel Moth, a senior project executive, has commented, that C++ AMP technology should help developers to used modern GPGPU techniques, which breaks programing barriers for code that uses hardware possibilities to the full. Software developers don’t need to use other compiler or learn new syntax. This provides considerable advantage in faster mastering of a new technology.