Instead of using intrinsics to take advantage of the NEON technology in your target processor, you can use a compiler that automatically vectorizes your standard C or C++ code. This can give access to NEON performance without requiring assembler or intrinsics programming.
Both GCC and RVCT support automatic vectorization for NEON technology, but because the C and C++ standards do not cover the concurrency aspects, you might have to provide the compiler with additional information to get full benefit. The required source code modifications are part of the standard language specifications, so they do not affect code portability between different platforms and architectures.
This section describes how to enable automatic vectorization for your compiler, and small modifications to your source code that can lead to great improvements in vectorized software performance.