type

The -fpu compiler option sets the math unit to be used for floating point arithmetic. If this option is not specified, the default is -fpu X87.

-fpu X87 will generate floating point instructions for the 387.

-fpu SSE will generate floating point instructions for SSE and SSE2 with some math support still done by the 387.

-fpu NEON will generate floating point instructions for ARMv7-A Advanced SIMD (Neon) extension. Specifying this option will automatically set the arch to at least armv7-a if not already set to a later processor.

Functions normally return a floating point value (Single or Double) in the st(0) register. Sometimes, this may be optimized by returning the value in the xmm0 register instead. This can be specified with Option("Sse") after the return type in a function's declaration or definition. Option("Sse") is ignored unless the source is compiled with the -fpu SSE command line option.