dumb optimization trick

I was looking at ways to save memory for object slots and I got it down to 84 bytes. What I found special about this number is that it's slightly less than 256/3. If I page align every 3rd object slot, direct page instruction will be 1 cycle faster for every 3rd object.

I was thinking of the same thing, actually... Another little trick is that I like to have it to where the end of a loop is when an index register is 0, so I don't need a cmp instruction.

Now I need to find a way to always have objects spawn into the page-aligned slots first so they have a higher chance of being in a page-aligned slot. Using a linked-list object slot system makes this more difficult.

I could probably take this one step further and have slow slot objects move into fast slots when a free fast slot becomes available.