More than 4 players

[Gabbe]

what bout GAE?

[silnarm]

what bout GAE?

Next version will support maps with 16 player slots, of which you will probably be allowed to use up to 12 in any one game.

[John.d.h]

what bout GAE?

Next version will support maps with 16 player slots, of which you will probably be allowed to use up to 12 in any one game.

Wow... sounds like some of us are going to need some hardware upgrades.

[ultifd]

Eh, we don't need something that will kill our comps.
Still cool, but only 10 percent of  the community will probably use all of the slots... 

[Fluffy203]

I'd use them and overclock your computers lol , i mean if they are out of date  this is the easiest game to run

[silnarm]

You are unlikely to need any hardware upgrades. I've run AI only games at 400 frames per second (ten times normal game speed) with only occasional 'slow downs' (to ~300-350). If you can run Glest with 4 players and no problems, I doubt GAE with 12 players will be a problem. I'll be doing plenty of testing before I decide for sure, so it may end up being only 8, but I suspect 12 will be fine.

The only potential performance issue currently is the changes for multi-platform, which has made distance calculations rather sluggish (integer square roots), but I plan to replace the 'calculation' with interpolated look-up tables, so this shan't be a problem by the time we get 0.3 ready.

The age old problem of too much stuff on the screen could of course be a problem, there are some tweaks for the Renderer coming, but probably not until after 0.3.

Cheers.

[Fluffy203]

The age old problem of too much stuff on the screen could of course be a problem, there are some tweaks for the Renderer coming, but probably not until after 0.3

Couldn't you incorporate a draw distance within the fog of war? i mean seems like it would be a really good thing , can't see it , don't render it 

[Gabbe]

I will definately use all the slots for LAN party!

[ultifd]

I see...we'll I guess we'll see...
still cool and somewhat amazing though I guess... 
 

[Gabbe]

12 different colours...

[ElimiNator]

Yah... Titi had a hard time...

[Gabbe]

16....

[@kukac@]

Beige, Sandstone, Gray and Granite, imagine these colours on a Guard. Are you sure you can make the difference between them?

[John.d.h]

Also, black wouldn't show up on the mini-map.

[Gabbe]

I know, i just found some and stuffed them up...the 4th google result....

[wciow]

16 player support!! thats gonna be some epic battles  

0.3 iwill be a huge release  

[Fluffy203]

i agree and i can't wait for it 

[titi]

Do you really think 16 is wise at the moment? Ok, maybe you got a fast pathfinder, but from what I see 8 is enough ( without lots of performance tuning ).  Try a 8 player game with CPU-Megas and after a while you will see what I'm talking about ( Even in single player games ).....
But maybe I'm wrong and you can handle it. Megaglest definitly still has serious problems with epic battles using more than 100 units on slow machines :-/ .  Even without battles it gets slow after a while with hundrets of units which have to be managed.

[Gabbe]

no! THIS MUST BE IMPLEMENTED! Our LAN is growing and we need capability to add all the people into a game! last time there were around 40 people, som played FPS and the MG was overpopulated...

[silnarm]

Do you really think 16 is wise at the moment?

I actually just said support for 16 player slots in maps, with a likely limit of 12 in the game.

Ok, maybe you got a fast pathfinder, but from what I see 8 is enough ( without lots of performance tuning ).

We have more than just a fast pathfinder Titi, we have a significantly streamlined unit update, and work in that area (and others) continues.  We've already done lots of performance tuning, and continue to.

But maybe I'm wrong and you can handle it. Megaglest definitly still has serious problems with epic battles using more than 100 units on slow machines :-/ .  Even without battles it gets slow after a while with hundrets of units which have to be managed.

We'll see.

I'll be doing plenty of testing before I decide for sure, so it may end up being only 8, but I suspect 12 will be fine.

Regarding performance, the unit update stuff is rather different to regular glest now, so that wouldn't be easy or straightforward to add to megaglest, but you may want to cast an eye on this commit from earlier today,
https://sourceforge.net/apps/trac/glestae/changeset/616

The interpolation stuff should be easy to lift out and put into megaglest, I don't think very much has changed in GAE's Model & InterpolationData classes.

I've only tested it on windows so far, where it is about twice as fast, by all accounts VC++ does a very ordinary job optimising SSE intrinsics, so the boost may be even more impressive under linux.

no! THIS MUST BE IMPLEMENTED! Our LAN is growing and we need capability to add all the people into a game! last time there were around 40 people, som played FPS and the MG was overpopulated...

With 16 slots in maps, changing the game itself to support 16 will be rather straight forward, maybe we'll make maxPlayers configurable, if you're at a LAN party, you're presumably talking about fairly powerful machines on average, so they may well handle 16 players without problem.

[wciow]

Another use for 16 players is for LUA scenarios. More factions allows more complexity in your scenarios.

I really don't see a problem with 16 player support. Why did MG bother to upgrade to 8 players? surely 4 was enough.

@Silnarm, I'm interested to know what you think the best quality vs size would be for model animations?

How good is Glests interpolation i.e. can we rely on Glest to get away with 4 frames per second and interpret them the same as if we have 8 frames per second in the model?

This is somewhat of a moot point since I can easily fit every model/texture in my Glest directory into my GPUs RAM and my entire Glest directory several times over into general RAM. 

[emscape]

why not load every animation once in the vram, instead of loading it 100 times if u use 100 swordsmen?

[Gabbe]

i guess dx and openGL will do that?

[ultifd]

Another use for 16 players is for LUA scenarios. More factions allows more complexity in your scenarios.
I really don't see a problem with 16 player support. Why did MG bother to upgrade to 8 players? surely 4 was enough.

Well, actually to me 8 is enough...but I can see for the future 16 or 12 would be nice too... 
We'll see... 

[silnarm]

@Silnarm, I'm interested to know what you think the best quality vs size would be for model animations?

How good is Glests interpolation i.e. can we rely on Glest to get away with 4 frames per second and interpret them the same as if we have 8 frames per second in the model?

This is somewhat of a moot point since I can easily fit every model/texture in my Glest directory into my GPUs RAM and my entire Glest directory several times over into general RAM.  

How good the interpolation is/seems depends on the motion in the animation, if you have parts of the model that move on a very curvy path, you'll need more frames than otherwise.  The interpolation is 'linear' between two key frames, reducing it to two dimensions, this is what is happening...



As to video ram requirements with this, they are the same as using no animation. All the mesh data is stored in system memory, the interpolation is done on the CPU and then the resultant mesh is sent to the video card, all those frames aren't put in video memory, so you don't need to worry about that.

Cheers.

[Edit: typos]