Lee said in his last blog: 'I am curious what the community (you guys and gals) would have me do at this point.'



I thought this was an interesting question and perhaps deserved a thread here.



For me, it still has to be about performance and stability. It's not sexy, but if any lessons can be learnt from classic FPSC, I think it's that piling new features on top of an outdated and somewhat flaky engine doesn't result in games people want to play.



Any thoughts?

well tbh I have yet to see any performance improvements, so to me it seems I need new hardware rather than improved software. so maybe a feature or 2 might be nice, and then some more performance. but I don't know what features he could add though, but then im no doctor so therefore have no patience lol.

I must admit, each update has got slightly slower for me. So performance is still top priority. I'm not quite sure what is causing it but the new update, although reporting a slightly better fps than before, seems laggier somehow. The pop up can be rather jarring at times also, but I imagine that will be fixed.

Performance is still the biggest problem I see. My computer can eat up any game I throw at it at maximum settings and 1920x1080 resolution, but FPSC:R often struggles to stay above 30 FPS even on the default river map which has very few assets.



However, I also know Lee is probably going insane trying to rid bottlenecks so it may be healthier for him to go on to some other tasks for awhile. What I do not know.

performance and stability! after that is done, then you can move on adding fpscx9 character and weapons support etc...

I just worry that after all the performance work lee has done so far, which as I said has not really made much difference to me sadly, that he adds 1 or 2 features then has to jump straight back onto performance for fixing it with the new features. so far its been 3 beta's with no meaningful effect on the software as I see it. I can look up at the skybox in the engine and get 110fps but it drops to 14fps when looking straight forwards, which then drops to 5fps when I add a patch of grass and a very little water, this is on the lowest settings so just how much can he really improve on I wonder without adding features can he maintain whatever his work has achieved. yes stability is important but does it not need a test to see if it can handle the features he desires aswell. it will all be for nothing if when he adds features he has to go back and find new ways to improve the performance and make it work for every1 yet again, just my opinion though lee is the 1 who has to do all of this stuff so.

yeah it would be fine if in the next beta the interface Gui is active.

or any little feature for some tests or so..

I have to agree it does seem to be a little slower I was getting in the 30 40 range now 24 there are many things that could be added, in beta 1.002 muti textures worked in 1.003 they don't, I think memory is going to be a problem too if you make a big level like a town as of right now mostly everybody is just doing block buildings in a wide open space to much time is being used on the plants and stuff and very little on real levels its great if all your games are in the jungle. I had a big level that a friend of mine who does work for a game company gave me and I tried to load 1/4 of that and it ran out of memory. so I did a couple of the buildings and the collision did not work right no body seems to address this. The other thing I see is if this program won't run on an off the shelf computer I see little hope of selling many games.

My experience is much the same as you guys. I'm just curious to know what particular feature of the code Lee's using is causing such low fps.



Edit Just to reinforce that point here's a screenshot of a demo DBPro scene containing 1000 Tree Magik trees. About 500 have been excluded in code as being obviously out of view and some others are presumably culled by the internal DBPro checks. DBPro is comfortably running this scene, with a shader on each tree, at just under 60 fps whereas with just a few objects FPSCR struggles to reach 10 fps. There isn't any terrain in my demo of course, just a flat plain, and my suspicion is that the terrain system is the problem with FPSCR.



I'll try to find time to see how my simple terrains fare with so many trees. It'll be a useful comparison.

that could be true, Green Grandalf



some time ago i attempt to make a nearly real forest.. no way maybe 50 low poly trees and the engine gone broken.



i have a Intel core i7, 8GB ram, Gforce GTX 570.

Did FPSC Classic suffer these same framerate issues? I ask because I've never used it--but I assume it's a new problem with Reloaded at which maybe it is the terrain itself. Wonder if Lee could try removing the terrain and replace with a single plain just to see how it affects performance.

Thanks for the feedback so far. It looks like I have more to do guys I am back in the office wednesday, lets see what happens if I switch off terrain in the next update

P.S. I am not Rick. I am Lee, on the road and using Rick's old phone

Trying to find out what exactly is causing this performance issue is high in my mind.



I'm still new to character side of game creation, will it be necessary for us to use a specific skeleton in the final released version, or do custom characters work fine as long as you have custom animation sets? A little advice on where I can be going with my modeling and content creation, in preparation for the final release would be great Lee. As I want to create lots of content over the next year before release of FPSCR reloaded, so I can create a finished game of high quality as quickly as possible. Obviously will have to wait until FPSCR final release is ready to create levels, which would be time consuming in itself.



Many thanks.

I don't really see why lots of content should slow things up - unless it's all in view at the same time. An interesting world to play doesn't necessarily have lots in view at any given moment. Depends how efficient the culling is I suppose.







The player will probably fall to oblivion in superfast time.



Seriously though, that would be an interesting comparison.

I tried this with nothing on the level the

pollys were 26655 fps 35 draw calls 68

then I went to the levels end and went over the edge and it was

pollys were 29 fps 207 draw calls 15 that's a lot of pollys for for a blank level

looking at that graph it almost full[/quote]Seriously though, that would be an interesting comparison.

It's the terrain which is gobbling up all those polygons. A travesty when you have a completely FLAT world I have improved the shaders at LOWEST level for lower end cards that cannot deal with 100K+ polygons but another optimization I am planning is to reduce the flat terrain scenario drastically, and hopefully make some of those redundant polygons go away. Please try BETA 1.004 which I think is the first 'proper' performance improving update so far.