Hexadecimal floating-point constants
The ARM compiler implements an extension to the syntax of numeric constants in C to enable explicit specification of floating-point constants as IEEE bit patterns.
The syntax for specifying floating-point constants as IEEE bit patterns is:
Interpret an 8-digit hex number
floatconstant. There must be exactly eight digits.
Interpret a 16-digit hex number
doubleconstant. There must be exactly 16 digits.