The overrun error usually occurs when your real-time model cannot execute a particular task or function within the specified sample time. The issue might be that the computations required by your Recurrent Neural Network (RNN) block exceed the processing time that is available per sample.

Here are a few suggestions to overcome the overrun issue:

Concerning the second part of your question, giving a block "more time" in a real-time system is generally impossible. In a real-time system, all operations are supposed to be completed within the specified time frame (the sample time). Therefore, if your RNN block needs more time than is available, you might have to consider one of the abovementioned options.

If your goal is to process a certain set of data and then skip a number of subsequent data, one way to implement this would be to use a counter inside your model that increments at each sample. You can then use a Switch block to decide whether to process or skip the current data based on the counter's value. Please note that this is not a common real-time control system strategy and might not apply in your situation. It also would not resolve the issue if your RNN block cannot process the data within the available sample time.