I find that video to be truly terrible misdirection - after all they compare their new best case scenario with the absolute worst case from the past. They acknowledge this, but who was genuinely doing this?
Put it like this, *currently* I can get roughly 30 fps animated 32x32 tiles on my 2nd gen ipod touch using my own code, of levels of any size (as in, I have the entire overland Zelda SNES map working fine), although speed ups are welcome, 'cos I am close enough to 30 fps that any game code slows it down!.
Where I think we’ll see the most significant speed boost are Lime levels that use physics, where objects off-screen must still be processed, but their visual side are culled by Corona.
Generally faster sprites? Sign me up!
Expecting genuinely amazing results? Not really… corona never really struck me as massively slow, but it boils down to the same problem that Graham has with Lime - people use it and corona in the laziest, quickest fashion possible, and naturally that always comes with performance hits.
Am looking forward to it, but I don’t expect it to revolutionise anything, since all you had to do to speed up tile engines was do your own culling anyway. [import]uid: 46639 topic_id: 12910 reply_id: 90342[/import]