With the synchronous update function all neurons change their value at the same time. All neurons calculate their activation in one single step. The output of all neurons will be calculated after the activation step. The difference to the serial_order update function is that the calculation of the output and activation value requires two progressions of all neurons. This kind of propagation is very useful for distributed systems (SIMD).