FPackedRGBA16N is divide by 32768 twice in NEON/FPU

FORCEINLINE VectorRegister4Float FPackedRGBA16N::GetVectorRegister() const
{
    VectorRegister4Float VectorToUnpack = VectorLoadSRGBA16N((void*)this);
    VectorToUnpack = VectorMultiply(VectorToUnpack, VectorSetFloat1(1.0f / 32767.0f));
    // Return unpacked vector register.
    return VectorToUnpack;
}

// FPU
FORCEINLINE VectorRegister4Float VectorLoadSRGBA16N(void* Ptr)
{
    float V[4];
    int16* E = (int16*)Ptr;
    V[0] = float(E[0]) / 32767.0f;
    V[1] = float(E[1]) / 32767.0f;
    V[2] = float(E[2]) / 32767.0f;
    V[3] = float(E[3]) / 32767.0f;
    return MakeVectorRegisterFloat(V[0], V[1], V[2], V[3]);
}

// NEON
FORCEINLINE VectorRegister4Float VectorLoadSRGBA16N(const void* Ptr)
{
    alignas(16) float V[4];
    int16* E = (int16*)Ptr;
    V[0] = float(E[0]);
    V[1] = float(E[1]);
    V[2] = float(E[2]);
    V[3] = float(E[3]);
    VectorRegister4Float Vec = VectorLoad(V);
    VectorRegister4Float Div = vdupq_n_f32(1.0f / 32767.0f);
    return VectorMultiply(Vec, Div);
}