Thank you for sharing your idea and your code and thoughts and thanks for your reply.

I would like to generalise it for sets of different sizes. Maybe it's something different to a gray codes which you taught me today, it feels that it should be simple.

Maybe someone kind can step in and help me or I can ask it on Stackoverflow.

I need to think it through.

At one point in time I was reading a section on Intel's website of cache optimisation using a tool that shows you the cache behaviour of the CPU and column or row major iteration it had a GUI. I found it very interesting but I cannot seem to find it at this time. I did find some medium article of Cachediff.