Continuing the strategy of making hardware by breaking the problem into increasingly tinier pieces until they become something I know how to do.
-
Continuing the strategy of making hardware by breaking the problem into increasingly tinier pieces until they become something I know how to do. As a result I... Think I'll have a fully functional tile renderer soon. Went all the way down to "expand a single byte into the right amount of pixel data", now wrapping up "render one row of a single tile" (which might span bytes and skip pixels), then on to "render a full scanline of tiles"
-
I had it working okay with less factoring, but this is easier to reason about, easier to test pieces in isolation, and it turns out easier to pipeline, so hopefully at the top level I can issue "gimme 640 pixels for scanline 123 and after a few cycles of spin-up I'll have a consistent one pixel per cycle, with some slack to allow for memory contention and the like.