Apart from AMD's announcement at AMD's Developer Summit 2013, of an implementation of C++ AMP targeting SPIR and HSAIL, they also announced that Bolt, a C++ Template library for GPU and multicore CPU programming, will have enhanced support for C++ AMP. Bolt, first announced at AFDS 2012, allows you to develop heterogeneous compute accelerated applications using pure C++. The library, currently at version 1.1, contains accelerated kernel code for many useful functions like Sort, Scan, Reduce and others, so you won't need to learn C++ AMP APIs in order to obtain the benefits of heterogeneous acceleration.
One of the cornerstone for C++ AMP is productivity (along with performance and portability) and we are happy to see libraries like Bolt natively use C++ AMP under the covers. We believe layering library functions on top of C++ AMP to improve performance is a great way to bring the benefits of heterogeneous computing to a broader set of developers.
The source to Bolt v1.1 is available on GitHub. For further details on Bolt features and capabilities along with Bolt architectural details and how-to information, visit AMD's Blog.