1.18 -- Go vislimit!
The page below describes the work-in-progress during 1.17. We finally delivered the doubled+ vis limit to players in version 1.18. It's less of a critical priority now, but we are still looking at ways to improve the update streams and will continue to do so...
1.17 -- Matilda Fleet
Well, we can certainly *display* more vehicles now...
With the unmodified 1.16 client, drawing all of these vehicles hits me for 30% of my frame rate (from 18 to 12 on my dev box running a debug build) - which might translate into a 15-20% loss for a player - or from 50 to 40 on my home box.
Moving them, firing their guns, etc, could take me as low as 2fps. We've made some pretty good progress there, but Martini and the other client developers had plenty of 1.17 tasks prescient upon their time during this experimental stage. We're still looking at a lot of 4-5 fps, and we obviously need to get that a long way up. At this point it's probably about 80% coding, just going through the code and finding hard coded assumptions and numbers, and 10% planning and about 10% of what I'd call "real work".
When do I get my increased vis limit?
Not yet. Raising the vis limit equates to increasing the bandwidth used by the client. Hard and simple. That doesn't mean we have to exclude 56k users, but with them included, we have a very tight, low budget ceiling for our bandwidth consumption. So we have to try and balance: bandwidth use vs smoothness of updates vs number of players visible vs bandwidth remaining for things like ordnance, voice, etc.
We also need to look hard at the framerate impact. More people visible equals more triangles. More people firing equals more effects, more physics, more polygons. More people moving around and firing equals more load on the sound sub-system.
Badger has also released some in-game images of the early, experimental client: click here.