Support for double-precision floating point basic arithmetic in ATI Stream SDK (2.1 and later)

To view the full article, access it here: KB88 - OpenCL™ Extension: Support for Double-Precision Floating Point Basic Arithmetic in OpenCL™ C Kernels

In short, before using double data types and/or double-precision floating point operators in OpenCL™ C kernels, you need to include the directive:

#pragma OPENCL EXTENSION cl_amd_fp64: enable