Lennart here, Both fundamental techniques were known beforehand, and in both cases the major innovation was adapting the algorithm for the specific compute pla

Lennart here, Both fundamental techniques were known beforehand, and in both cases the major innovation was adapting the algorithm for the specific compute platform. For us it was exploiting FPGAs to compute these P-Coëfficients, for Jäkel, his major contribution was restructuring the summation as large matrix multiplications, which could then be executed very efficiently on GPU.