The inline assembler provides direct support for VFPv2 instructions. For example:

float foo(float f, float g)
{
  float h;
    __asm
    {
        VADD h, f, 0.5*g; // h = f + 0.5*g
    }

    return h;
}

In inline assembler you cannot use the VFP instruction VMOV to transfer between an ARM register and half of a doubleword extension register. Instead, you can use the instruction VMOV to transfer between an ARM register and a single-precision VFP register.

If you change the FPSCR register using inline assembler, it produces runtime effects on the inline VFP code and on subsequent compiler-generated VFP code.