xendance

Are the memory points for the A-164 rockets/missiles configured properly? plane selectionPosition "Rocket_1" and plane selectionPosition "Rocket_2" both return [0,0,0] when plane is an A-164. But with Neophron plane selectionPosition "Rocket_1" returns [-4.3218,-0.220851,1.06955] and for Rocket_2 it returns [4.28913,-0.220851,1.06955].

I'm looking for testers, mainly people who are making their own aircrafts since this script is supposed to be kind of a bootstrap script that the modder can use the implement CCIP. Send me a PM if you're interested

Could you add the actual mathematical formula to each function in the wiki? That way we could verify the behavior of the script commands if they break in the future for some reason.

Shame on BI for having automated testing!

There has to be some sort of visual feedback (at least for AI), maybe in a form of animation where he quickly looks/checks if he was hit and where. Or something along those lines.

You're just flat out wrong. Of course you'll get similar death poses when you shoot the skeleton in the same place with the same weapon while they're in the same pose. :rolleyes:

Should be rather trivial to do a trace from the formation start to formation end and see if it intersects with anything. And the vehicle commander is the one who tells them to which side to dismount, if he doesn't know that there are enemies on the left side and says "dismount to the left"... well, that's just realistic. If it becomes a problem, commander enemy detection should be tweaked rather than the dismount procedure.

This is really bothering me. It seems such a simple fix, don't you have interns for this kind of stuff? :P

As I said previously: The game is mostly single threaded, you won't see higher CPU usage than one physical core utilized at 100%.

The game and server is is mainly single threaded. You probably have hyper threading on, so 0.27 * 8 logical cores = 1 physical core maxed out. So, to answer your question on "what limits the server?": It is limited by the single thread performance of your CPU.

That's why you should attach the objects to the turret, not the body :P

Devs have the source code. Devs have the profiling tools. Devs know their engine inside out. I really, really honestly doubt this is the cause and I really doubt devs don't know why it performs poorly.

Oh, you mean the stencil shadows? Why did you then link this and this? Those are examples of light mapping. You said "baking" too, which lead me to think you were talking about light mapping. As I mentioned, shadow LODs are used for stencil shadows which are 100% dynamic. They aren't baked and the links you posted have nothing to do with stencil shadowing. Neither does baking. Regarding the memory issue: As I previously said, each lightmap requires certain amount of space to look even remotely good. Due to the amount of objects on Altis, you'd end up needing gigabytes of video memory.

"Shadow polygons"? Shadow LOD? What? You mean the lightmap UVs or what? I wasn't talking about processing power anyway, I was talking about memory usage.

In ArmA 3 for example every tree shares the same normal map and ambient occlusion textures, but baked lighting would require a unique lightmap texture for every single tree. The memory requirements for that would be humongous, even more than the island satellite maps.

What on earth is this magical culling that supposedly is easier to do than just render shadowmaps?

I don't plan on doing the HUD part. What you see in the videos is just the debug view (rockets even had the trajectory debug lines visible) with the x-hair. The plan is to release the script for everyone to provide modders a CCIP solution that they can use with their vehicles. That way they can handle the HUD themselves and do what they want with the CCIP position.

Cheers, I'll take a look at the guide.

It's bright for a reason. If a person is playing without shadows a soldier inside a pitch black building will literally glow. Now that they aren't too dark, the soldiers blend with the overall lighting better.

I've been working on a CCIP script for jets. I use Euler's method to approximate the flight path of the projectile which works nicely. The actual process goes like: Determine the velocity for each velocity vector: v ↠v + Δv = v + a*Δt Determine the position for each axis: pos ↠pos + Δpos = pos + v*Δt Rinse and repeat until we hit ground. The problem is that I don't know how air friction is applied to the equation. I've been also puzzled by the fact that missiles/rockets have positive airFriction config, while bullet type projectiles have negative. Does anyone know how to apply the air friction to the equation?

I think I managed to reverse engineer it. Red line was done with the projectile path tracing script and the dashed line is my script's estimation (dashes represent the time steps). It seems close enough to me: http://cloud-3.steampowered.com/ugc/576766755283217560/F7E86F86F67624CF1C9E9E5E4E4708DAB31231D9/ http://cloud-2.steampowered.com/ugc/576766755283219098/85E8FB045124D2E8103E3B0EB6F73B885B8C58D4/ http://cloud-3.steampowered.com/ugc/576766755283220122/DEBAAE9C8992C7EF6110E7B386C325C5A1D0546D/ Next up: do the same for bombs and double check everything works while flying.

Actually, I'm not sure if I'm calculating the rocket thrust ok... Here's the code I'm using (I snipped some unnecessary portions off): //total velocity accounts for aircraft speed _vx = _totalVelocityX; _vy = _totalVelocityY; _vz = _totalVelocityZ; //_pos is aircraft position _posX = _pos select 0; _posY = _pos select 1; _posZ = _pos select 2; //590 is just the max speed the rocket has, actual thrust/mass doesn't give enough acceleration _accelerationV = [_gunUnitV, 590]call CBA_fnc_scaleVectTo; for "_i" from 1 to _maxIterations do { private["_ax","_ay","_az"]; //_dt = _maxResolution min (_i^-1); //_dt = _dt max _minResolution; _dt = GLOBAL_DT; _elapsedTime = _elapsedTime + _dt; _ax = 0; _ay = 0; _az = -_gravity; if(_elapsedTime < _thrustTTL) then { _ax = _ax + (_accelerationV select 0); _ay = _ay + (_accelerationV select 1); _az = _az + (_accelerationV select 2); }; //http://physics.gmu.edu/~amin/phys251/Topics/NumAnalysis/Odes/projectileMotion.html _Vmagnitude = sqrt(_vx^2 + _vy^2 + _vz^2); //disable this for now _FdragX = 0;//_airFriction * _Vmagnitude * _vx; _FdragY = 0;//_airFriction * _Vmagnitude * _vy; _FdragZ = 0;//_airFriction * _Vmagnitude * _vz; //determine the velocity: v ↠v + Δv = v + a*Δt _vx = _vx + (_ax - _FdragX) * _dt; _vy = _vy + (_ay - _FdragY) * _dt; _vz = _vz + (_az - _FdragZ) * _dt; //determine the position: pos ↠pos + Δpos = pos + v*Δt _posX = _posX + _vx * _dt; _posY = _posY + _vy * _dt; _posZ = _posZ + _vz * _dt; } And that gives me following result: http://cloud-2.steampowered.com/ugc/579018011189692481/C9CB8AF35EBD4D9E162A7867B4DCE625AB57B710/ http://cloud-3.steampowered.com/ugc/579018011189694649/EB5FAAC1979250D9FC4A45F338C6254E3312082D/ You can see that the estimation is way short, and would be even more short if I applied air friction... So I suppose my new question is: How is the rocket/missile acceleration calculated in the engine?

Thanks but I'm looking for the actual equation. As in how does the air friction coefficient affect the acceleration of an object, because it clearly does affect rockets.

Yea I did the same. Unfortunately now my layer generation takes over three hours instead of few minutes :( I would like some general info about how the landgrid affects performance. It's fairly obvious that it relates to the subdivision of the spatial data structures that contain the drawable objects. As in does a high landgrid count require more CPU time? Related to this: http://feedback.arma3.com/view.php?id=18158

Yeah, I found this half a year ago, the heightmap quality is great.