You copied the Doc URL to your clipboard.

Compiler and library support for half-precision floating-point numbers

Half-precision is a floating-point format that occupies 16 bits.

Half-precision floating-point numbers are provided by:

  • The Vector Floating-Point (VFP) Version 4 architecture.

  • An optional extension to the VFPv3 architecture.

If a VFP coprocessor is not available, or if a VFPv3 coprocessor is used that does not have the extension, half-precision floating-point numbers are supported through the floating-point library fplib.

Half-precision floating-point numbers can only be used when selected with the --fp16_format=format compiler command-line option.

The C++ name mangling for the half-precision data type is specified in the C++ generic Application Binary Interface (ABI).