3 ms·

I am curious how does this compare to nvidia's rapids.ai / cudf.