Largest Number in an Array

A count sits at 2000H, and that many bytes follow from 2001H. Find the largest one.

You can already walk an array and accumulate a total. This problem changes what happens inside the loop: instead of combining every element into the answer, you choose between them. That needs a comparison, and the 8085 has no instruction that says "is this bigger". What it has is subtraction — and the observation that once you have subtracted, the flags already contain the answer, so the difference itself can be thrown away.

Two things go wrong for almost everybody the first time:

• Comparing with the wrong flag. The sign flag looks like it should mean "smaller", and for the data most people test by hand it appears to work.
• Starting the running best at zero. It works here, by luck, and stops working the moment the same program is asked for the smallest element.

The first step is a comparison with no loop at all. The second puts it inside one. The third asks for the smallest element as well, in the same single pass — which is where starting from zero stops being a matter of taste and starts being a wrong answer.