difference among static_cast<>, (int), and __float2int_rn() in cuda

Question

I have a question about the difference between type conversions in CUDA:

static_cast(1.3f)
(int)1.3f
__float2int_rn(1.3f)

For