Hi All,
I have one general doubt, why do simulation runs for hours even though the simulation time in the run times are in nana seconds and are meant for running for some nanoseconds. I am pursuing masters in VLSI, sorry for noob questions asked.
Thank You
In reply to skyhome0911:
I think clay animation is a good analogy to logic simulation.
Suppose we want to create a movie of a clay figure walking down the street. That figure represents our “model” of person in clay. Imagine the number of steps you have to do for each 1 second frame of the movie. You have to move yourself into the frame, adjust several body parts of the figure, move out of the frame, then take a picture. If your are really skilled at it, maybe you can do one frame every 5 seconds.
But now try to create a movie of a busy city street with 100s of people, cars, etc. that need animation. It’s going to take hours for that one person to make all the adjustments for one frame. You could add more people to help, but not so much that they start getting in the way of each other. You could also take shortcuts by not providing as many detailed movements for subjects not in the focus of the frame.
Now back to VLSI (Very Large Scale Integration). We are trying to animate millions, and now billions of devices running concurrently in software where each step in simulation is at best taking a single CPU instruction cycle. Of course there are always optimizations and higher abstraction levels that can improve performance to some degree.