https://de.mathworks.com/matlabcentral/answers/48239-how-to-do-area-optimization-like-sharing-and-streaming-in-models-with-feedback-simulink-hdl-coder
The answer given in the question on top is also working for Speed optimizations(adaptive Pipelining): "consider moving the feedback loop into the subsystem where sharing is occuring".
I did that but my mistake was that the discrete Integrator, which is build with a "sum" and an "UnitDelay", also results in a Feedback path/Loop.