I have a question about simulation speed and performance - what's the tradeoff between driving input signals in a module and forcing them using tcl commands?
Is forcing using tcl slower compared to driving from a testbench? I'm guessing it is if the tcl script gets complex? And is the answer the same for questasim and vivado simulator?
The points I would make are...
A testbench can be tightly integrated with the simulation. (e.g. triggered on the clock). It is also automatically trigger and run and is repeatable. Unless you are doing a one off, I would always use the testbench,
Hope that helps