You copied the Doc URL to your clipboard.

Reserved data types

List of reserved data types. Do not use these in your OpenCL kernel code.

Table A-4 Reserved data types

OpenCL Type Description
booln Boolean vector.
halfn 16-bit float, vector.
quad, quadn 128-bit float, scalar, and vector.
complex half, complex halfn Complex 16-bit float, scalar, and vector.
imaginary half, imaginary halfn Imaginary 16-bit complex, scalar, and vector.
complex float, complex floatn, Complex 32-bit float, scalar, and vector.
imaginary float, imaginary floatn Imaginary 32-bit float, scalar, and vector.
complex double, complex doublen Complex 64-bit float, scalar, and vector.
imaginary double, imaginary doublen Imaginary 64-bit float, scalar, and vector.
complex quad, complex quadn Complex 128-bit float, scalar, and vector.
imaginary quad, imaginary quadn Imaginary 128-bit float, scalar, and vector.
floatnxm n*m matrix of 32-bit floats.
doublenxm n*m matrix of 64-bit floats.
long double, long doublen 64-bit - 128-bit float, scalar, and vector.
long long, long longnb 128-bit signed int, scalar, and vector.
unsigned long long, ulong long, ulonglongn 128-bit unsigned int, scalar, and vector.

Note

  • The half and half vector data types can be used with the cl_khr_fp16 extension.
  • The double and double vector data types can be used with the cl_khr_fp64 extension on Mali™ Midgard GPUs.
Was this page helpful? Yes No