i'm able to get 0.8x amplification between the first and the third core. this is *less* than the amplification i got when cascading only one core of the first, so not likely a good direction to pursue, though i haven't yet explored that much of the parameter space.