Comparing "nextgen" engines stuff with BMax default rendering is fun, but it doesn't tell anything. While a "nextgen" engine won't run on most onboard GPUs and older cards (and ATI :D), most of us want to see their stuff run on as most machines as possible.
If I make a particle system for my GPU I don't care about rendering 1000 unseen particles. Hell, I even could pixel them with Plot and get 80fps. And guess it, updating my particles with a single call raise my fps from 900 to 1200. Unfortunately most onboard gpu's I've tested this with don't like buffered output and run slower with this method.
So, I optimize my engine to run as fast as possible on default office PCs with onboard Intel graphics. That's the minimum spec most of us want to serve, so that's the spec to optimize everything against. Anything wrong with this?
If I make a particle system for my GPU I don't care about rendering 1000 unseen particles. Hell, I even could pixel them with Plot and get 80fps. And guess it, updating my particles with a single call raise my fps from 900 to 1200. Unfortunately most onboard gpu's I've tested this with don't like buffered output and run slower with this method.
So, I optimize my engine to run as fast as possible on default office PCs with onboard Intel graphics. That's the minimum spec most of us want to serve, so that's the spec to optimize everything against. Anything wrong with this?