[monocasa]

You know how you can't just compile normal C to a GPU and expect a speed up because the underlying computational models are so different? It's like that, just a few orders of magnitude worse.