Not running C directly, but getting about 200,000 iterations per second running through an Nvidia GPU through CUDA and the Python wrapper PyCUDA.