Tug of war!

In class, we will be discussing a race in which one thread tries to increment a variable while the other tries to decrease it. Let's say we start with the variable at n. We declare victory if the variable ever gets to 0 or 2n.

1.

Modify your program so that it conducts this race. Of course, if n is small, you would expect whichever process runs first to win. But if it takes more than one quantum to iterate n times, then there might actually be some back and forth. In fact, if the CPU scheduler strictly alternates, the program may never terminate.

2.

Try out some different values of n and see how long the program takes. I am not sure whether this will work or not, but it is possible that we will be able to estimate the quantum this way (or at least guess what the order of magnitude is).

3.

Try this experiment on iris12, too. How do you expect the behavior to differ if the two threads are running on different processors?