5 ms·

A comparison with the C256 Foenix Gen X would also be useful.

• cmrdporcupine 4y ago
  Back when I still had an order in for the Foenix I worked on an emulator (in C++) for a bit. It's likely changed a lot since then -- so take this with a grain of salt -- but it was a lot less "VIC"-like than this and more like a souped-up SVGA. Its framebuffer was structured like an 8-bit indexed palette linear memory. And then she also offered tiling & sprites and she was also working on a blitter at the time. It was quite nicely and thoughtfully structured though my biggest beef was that writing code for the '816 for managing that framebuffer was a bit of a hassle as you have to fiddle with the bank register to work with pointers into regions bigger than 64k. So let's say you're writing a routine to flood fill or draw lines or rectangles etc, you would, I think, need to be managing the bank register for all framebuffer addressing, which is going to add cycles and complexity. If I were designing a system around the '816 (or any other processor without a large linear address space) with a large framebuffer like that I'd consider having some planar graphics modes where each plane doesn't exceed 64k in size. Then at least you could do single-colour operations quickly/easily. And it sounds like the MEGA65's VIC derivative is more 8/16-bit friendly around these kinds of issues.