**Joshua S, Y10**

The Collatz Sequence starts with any positive number and either halves it if it is even or triples it and adds one if it is odd, continuing until it reaches 1 after which it stops. The unproven theory says that for all positive integers this sequence should eventually reach one.

When I first looked at the collatz sequence problem for purposes of assembly language, the first thing I did was decompose the required algorithm into its component parts; an even/odd checker, a part that would triple a number and add 1, a part that would halve a number and a part that would stop the algorithm once one was reached. I then realised that it was nowhere near as hard as it seemed – each of the four main elements were fairly simple, the hardest being the division. Mr Gwilt demonstrated the method to achieve the odd/even checker in class. The triple and add one was easy – simply add the number to itself twice and add one, although this exact code isn’t demonstrated in my code below, as I also integrated the stop code into the odd function*. The division, however, required a more complicated algorithm, and two variables. It worked by loading the value of x and storing it as y, then looping code that would load y, subtract one, and store the new value. It would then undergo a check in which y was doubled and then x subtracted – if y was half of x, the program would divert to another label called output, which would output x and return to loop, else it would continue looping until it got there.

* by subtracting one from the accumulator (which I had just loaded with the current value of x, the variable used to hold the current term) then using the BRZ command, I could check if x was currently 1, then divert to the stop code if it was. Since I had subtracted one, I had to add two to get the triple-add-one effect, and store a new x.

You can try out Little Man Computer (by Peter Higginson) online here: