You could try running the Performance Advisor to help locate the bottleneck. https://www.mathworks.com/help/simulink/ug/consult-the-performance-advisor.html
If your model is very large, you could also convert the larger subsystems to model references. MOdel references allow Simulink to incrementally update and generate code for each sub-model rather than trying to generate code for the entire large model at one time. An added benefit of using referenced models is that if you have a multicore computer, Simulink can generate code for child models in parallel on each of the cores ( requires Parallel Computing Toolbox ).
