Branch predictions

I spent most of the day today working on my project for computer architecture. Our assignment consists of programming a simulator for branch predictions. Actually, we are supposed to program three different predictors: one that looks at global history, another one that monitors local branch history results, and a third one that considers the predictions from the first two methods and makes a choice based on those and the accuracy history of each of those.

I had the choice between programming in Java, C and C++ and (much to my surprise) I decided to do the coding in C. (I guess that means I'm over the aversion that I developed during our Nachos projects last quarter.) I coded the first part before lunch and the two latter parts after lunch, only once got a segmentation fault(!) and then fixed that one without headache.

After work we went to the gym, where I added 10 pounds to the weights I was lifting and raised the bottom of the bench where I did my sit-ups up two notches. Now, who would have predicted that ...