5.7.1.11 Hard-float Option

The -mhard-float option indicates to the compiler that floating-point calculations in generated code should be performed by the hardware floating-point unit (FPU). A varient of the library that utilises hardware floating-point instructions is also selected for link time.

This option does not typically need to be specified, as the compiler will use an appropriate library and generate suitable floating-point code based on information provided by the Device Family Pack associated with the selected device.