Comment by jdw64

You really explain things in a way that's very easy to understand. I think I finally get it now. I understood the conditional branch part, but I think I was having trouble with the indirect branch predictor. So basically, when a branch instruction is executed, the CPU learns the actual target address. In the case of a switch statement, because the key is a single value like s, the history records get mixed up, whereas with a computed goto, since there are multiple keys, they don't get mixed up. Thank you for the clear explanation. It's a bit embarrassing that I didn't know this, but thanks to people like you, I feel like I'm learning more.