garbage script??

[WildFire]

:helpsmilie:I was wondering if anyone has anything for use as a garbage script? I have a rather large mission Ive been testing the last week and it is really fun, about 60+ pieces of armor take a friendly airfield that is somewhat defended. It gets relatively insane with all the AAA and small arms going back and forth.

 

The game runs relatively smooth on my machine, however it lag spikes with different players. I know I have one option, limit the number of advancing tanks. But of course we can see how that would limit the fun and replay-ability of this mission.:protest: Instead I would like to attempt to limit the graphics.

 

Does anyone know how to clean up smoking vehicles or have some kind of garbage collecting script? Is it possible to add a trigger for each vehicle once dead>deactivate=no more smoke?

 

Going to test today with scenes set to low, traffic off, and anything else I can limit using triggers. Thanks for any help.

[Speed]

Back in 2009 I tried to make a Black Shark mission with hundreds and hundreds of enemy soldiers attacking a farp. Turns out that dead units cannot be deactivated and made to disappear- though there may be a way through Lua to do this, I have yet to find anything in the five Lua environments I'm aware of that would allow it. That said, there's a lot of functions in the server environment (not to be confused with the net environment where, for example, Servman exists if installed) that I haven't figured out the use for yet.

 

I'll probably be releasing a scripting guide in the next 3 weeks-2 months that will detail the different Lua environments DCS uses, some useful functions in them, and how to access them. Originally, I was just going to do a scripting guide based in the server environment, but I'm convinced I need to expand it to cover the other four environments as well (the five environments, as named by ED, are: server, net, mission, export, and config).

Edited July 7, 2011 by Speed

[snwboardn]

I'll probably be releasing a scripting guide in the next 3 weeks-2 months that will detail the different Lua environments DCS uses, some useful functions in them, and how to access them. Originally, I was just going to do a scripting guide based in the server environment, but I'm convinced I need to expand it to cover the other four environments as well (the five environments, as named by ED, are: server, net, mission, export, and config).

 

We definitely appreciate the time you are putting into this.

[WildFire]

I'm glad you responded speed, I know if you say it cant be done we're screwed. Im looking forward to that guide, as of now I dont know where to find the functions or environments of this engine, and would easily read this just to be able to figure my way out around lua.

 

I've noticed in the dcs log that smoke "effects" die and then get recreated. Really I just need a way to end the loop as I see it seems to keep running the cycle indefinitely. Maybe there is a time limit but I havent run the server long enough to find out, nor would I benefit from it here. But if it does have a timer of some kind of counter perhaps we can change some variable just for one mission.... Hopefully.

 

Anyhow, thanks for your input.

[Speed]

I'm glad you responded speed, I know if you say it cant be done we're screwed. Im looking forward to that guide, as of now I dont know where to find the functions or environments of this engine, and would easily read this just to be able to figure my way out around lua.

 

I've noticed in the dcs log that smoke "effects" die and then get recreated. Really I just need a way to end the loop as I see it seems to keep running the cycle indefinitely. Maybe there is a time limit but I havent run the server long enough to find out, nor would I benefit from it here. But if it does have a timer of some kind of counter perhaps we can change some variable just for one mission.... Hopefully.

 

Anyhow, thanks for your input.

 

I didn't say it couldn't be done... I said I don't how to do it, and I haven't found anything that would allow me to remove dead units yet. There's a big difference between that and saying it can't be done... I don't know nearly enough about the program or even programming to say whether or not it can be done.

 

As far as smoke effects dying and being recreated... there is in fact a burn time variable for destroyable objects, and I think you can access this through Lua. I forgot where exactly it is, maybe it's accessible in the "server" environment, since the server environment stores a vast database of object information. Unfortunately, it's probably loaded into the mission at mission start and not accessed again, just like the mission table in the "mission" environment. See, as an example, you can access the unzipped mission table in the "mission" Lua environment, and do all the changes you want- like changing waypoints- but it doesn't make any effect in the world because that data was all loaded into memory somewhere else by the exe right at mission start and will not be accessed again by the exe. There may be ways to force the exe re-read in the data from the Lua mission table stored in memory, but if so, I donno how exactly- and I've already tried a few things to try to force this.

 

BTW, Lua does not have the capability to use pointers (pointers allow you to set or get data based solely on the variable's address in memory). Kinda sucks... but I don't think Lua was intended for that level of programming.

 

Anyway, in the most likely case that changing the burn time that is stored in Lua does nothing, maybe there is a db.reload function or something like that, so maybe you could change the value of the burn time, then force the exe to re-read the database. All this said, this change would most likely only work for the host.

 

Anyway though, if you just want to remove smoke effects and make vehicles only burn for like 30 seconds, there is in fact a Lua file you can modify. What I'm saying is I donno how you could go about changing that burn time through a pure memory accesses.

Edited July 7, 2011 by Speed

[WildFire]

I didn't say it couldn't be done... I said I don't how to do it, and I haven't found anything that would allow me to remove dead units yet. There's a big difference between that and saying it can't be done... I don't know nearly enough about the program or even programming to say whether or not it can be done.

That was kind of the point. I was happy you didnt say "it cant be done". I was looking for anything other than that.

 

 

Anyway though, if you just want to remove smoke effects and make vehicles only burn for like 30 seconds, there is in fact a Lua file you can modify. What I'm saying is I donno how you could go about changing that burn time through a pure memory accesses.

Would this only affect the host? Im intrigued about this file you said I could modify, maybe I could start looking there..... Where?

[Speed]

They are all in C:\Program Files\Eagle Dynamics\DCS A-10C\Scripts\Database\vehicles in folders.

 

All of those Lua files in there will probably have the GT.visual.fire_time variable. Change that and let us know what happens. Get Notepad++ if you don't have it.

 

That variable is also accessible through Lua in the server environment while the game is running, I just checked, but like I said, I'm guessing that changing it will not do anything. But I'm not sure.

Edited July 7, 2011 by Speed

[WildFire]

Yeah I downloaded N++ so I could read and learn your chatIO... which ive used in every mission since... Thanks I'll take a look and let everyone know. See what I cant effectively screw up...:evil:

[WildFire]

Wow, may have found something, take a quick look at config/world/world.lua

Seems to have burn times associated with each vehicle. I see also what you were talking about how each vehicle is given its own burn time variable(the GM variables). I need something more along the lines of global variables. But insofar, this looks promising.

 

The only problem so far is that these are burn times and I dont think that means smoke effects. So I'll continue looking. Neat stuff anyhow, perhaps we could have ED modify these files in the future so that smoke effects wont last forever and everyone will be able to make larger battles.

 

Oh and found a smoke folder, what program are we supposed to use to read .cfg's, or is it even readable...?

Nevermind looks like notepad works just fine... der.

Edited July 9, 2011 by WildFire
nevermind..

[Druid_]

Don't use notepad, it will cause problems!

 

Download Notepad++

[WildFire]

Yeah I was talking about N++ i just didnt put the ++'s in there. Thanks tho, I know. Turns out Speed of course was right only the GT variables change the actual fire time, those conflagration time things dont do anything as far as I saw. Of course those GT variables only affect client side, did a test with someone last night and it didnt affect his graphics on my mission. So I'm thinking maybe I'll put together a modman package and lower all the smoke times. Im going to make a main thread looking for input and see if this would be something the community might back, or want.

 

Right now it takes a T-80 20 minutes for the smoke to dissipate, I know some guys that said the highway in hell in Iraq took up to two days to stop smoldering so I understand the compromise ED is trying to make having long smoke times but it seems tough to compromise between graphic realism and limiting graphics for the purpose of multiplayer...

 

I have yet to find the fire control for buildings, does anyone know where to find that? I have been looking in every file I cant seem to find it.

[Druid_]

Are you sure that its the smoke effects causing you frame rate problems? Most modern graphics cards can render smoke with little FPS loss if done correctly IIRC.

[WildFire]

Well the problem I dont think can be limited to one thing actually. Secondly they arent my problems, when I run the mission on the server it runs very smooth for me the problem is other players reporting lower framerates. Ultimately I am actually intending to find a way to make missions larger, it isnt so much as I'm trying to fix a single mission as I am trying to open a doorway to make this game huge, for everyone.

 

My visions are kind of large, you ever remember flying around in F4 and you could fly across the FLOT and there was a living breathing war down there, watching it play out from above was almost magical, something I thought games would never be able to do.

 

Now we finally have the updated aircraft, and more to come but the battles suffer. I have yet to play a mission where I feel like I'm in the middle of a war... it always just feels like an isolated escalation. And I know we, (we as in the community, or the developers) can do better. This is not to downplay any of the work the developers have already done and my signature is a testament to that, however, I feel like weve got a lot of potential here.

 

So necessarily the smoke is not the only problem, the primary problem of course is having a 50+ units on the battlefield all active of course. What I am aiming and intending to solve is how to limit the graphics and any other effects/additions that can be limited in order to open the door to a much larger world.

 

This begins with limiting special effects that run, in example like a burning T-80(or any of the BMP's) that does so for 20 mins. Not a problem when you have 1 tank, but 60, obviously you can see how GPU usage goes exponential. Anything we can limit, or streamline to give mission makers the edge, is really what I'm going for.

 

Really I could drop the whole idea and stick with making missions that have a few quick tasks a few pieces of armor to shoot at same old same old... No problem. In my opinion so many people have mentioned wanting something like the dynamic campaign for falcon (which of course is not possible) but why do we want it? Is it because the AI did stuff and was active, or was it because you felt like you were part of something bigger? Part of a real war? I may not be able to make a dynamic campaign, but wouldnt it be cool if every mission we made, made you feel like you were in the middle of something huge? In the middle of a real war?

 

Who knows maybe im just crazy and no one actually wants this... very possible....

[Luse]

F4 had incredible Internal systems modeling for it's time.

 

It's graphics as everybody knows, Are subpar for todays simulation world, With all the complex equations and problems your processor is doing, ALONG with rendering the High-Def poly models that ED is putting into sims these days. It's a sacrifice between looking good and having a world within a world so to speak.

 

Also, F4 models weren't exactly uhh... the best around. :thumbup:

[WildFire]

Exactly right, now we got all these hi-def graphics and I'm trying to limit that so I can add to the gameplay, it is definitely a sacrifice. Honestly F4 isnt that great a comparison, but unless Im gonna compare A-10 to fsx, which is laughable, its the only somewhat legit comparison I have.

[snwboardn]

My visions are kind of large, you ever remember flying around in F4 and you could fly across the FLOT and there was a living breathing war down there, watching it play out from above was almost magical, something I thought games would never be able to do.

 

I was thinking the same thing but on a different lines actually... We can't have a dynamic campaign that ends on every mission and goes to a new one. How about we just make it an ongoing campaign that is basically one mission? My thoughts were having the player refuel and rearm and then get new orders, new waypoints and be briefed in the cockpit.

 

What you are trying to accomplish would go a long way with that.

 

Of course with my idea you wouldn't be able to save your progress, but just one mission that felt alive, would go a long way.

 

The day strategy and sims finally merge will be the day that I will be a happy man.

[WildFire]

How about we just make it an ongoing campaign that is basically one mission? My thoughts were having the player refuel and rearm and then get new orders, new waypoints and be briefed in the cockpit.

This actually can already be done. All you have to do is set up triggers to detect when the player has landed, or refueled or some other parameter and set up a few flags to go off and using either the editor messages or speeds chatIO(only in multiplayer) scripts you can broadcast messages. Im pretty sure you can load multiple flightplans in the cdu, I used to work on them in the navy and I know they are capable of doing so just a question of how they have been programmed to be used in the A-10, I dont know. In fact the only thing I think stopping people from making missions like this is because:

 

1.most people playing multiplayer rarely play longer than a couple hours on a single server

2.It takes a bunch of flags, and a bunch of triggers that can be a pain in the *** to get to work with each other correctly and to add all the conditions to each client in MP is a real pain in the neck

3.Most people want to get in the air and kill stuff, and there is a select few that would rather just spawn in a new plane to get refueled/rearmed than actually RTB and do the whole thing. It tends to take precious time that some people just dont have. So it throws the whole mission plan out the window.

 

I personally agree and tend to side with realism so I'll go back to base, but watching a bunch of people play when I have my server up testing missions, I dont see a whole lot that tend to do the same, although I am positive there are far more out there.

 

The real point of my thread here is that regardless of mission time or size that you get in the air and all he** is breaking loose. I want the battlefield to be big, to be crazy. I want to be able to look out the window and see a war. As it stands now, this doesnt seem possible, at least not on the scale that I hypothesize is possible.

 

One mission I have is called HammerDown, it has a tasking system with 7 tasks, randomly selected so the order is constantly different. There are several convoys that move all the way across the map and when they get chosen they could be pretty far or pretty close to the main area, and different convoys are spawned each mission. There are also 2 JTAC tasks of which 8 locations may be randomly generated. 4 different base locations for the random selection of the base task. There are also 30 SAMs of almost everytype that are spawned randomly, sometimes you get 3 sometimes you get 10.... That mission will be released soon so anyone can take a look at it. Basically it is pretty tough to memorize where stuff will be on the map, which was my ultimate goal. You can play it ten times and never get anything close to all the same pieces.

 

This was my first MP mission so you can obviously see building a large mission is not a problem, building a large battlefield is the ultimate goal here.

[snwboardn]

Well when you feel comfortable putting it out there I will definitely take a look at it. I am kind of holding off on my mission building right now because I want delayed take offs, for the AI, so bad it is killing my motivation to move forward with other things.

 

Once the new patch comes out and allows me to put idle AI on the field, I am going to see if I can pack the airports full of AI and have them scramble.

[Druid_]

I have a CAS mission which will take you over 4 hours to complete if everything goes well. I must admit with 180 triggers its starting to cause a few FPS problems so I might have to hack it into pieces.

 

I have however through scripting managed to add a LOAD & SAVE option so you dont have to restart the whole mission from the begining each time. Now it not Dynamic as previously destroyed units won't appear again in their destroyed position smoking away but its not really required for this type of CAS mission as I have you moving around the map supporting ground units.

 

I understand Wildfires passion for creating a large scale battlefield environment but to be honest when tasked you would only be dealing with a small part of it IRL. Having multiple units fight it out even 80nm away from your local battle is only going to use up unecessary CPU utilisation.

 

DCS just isn't designed to handle large scale battles so for now as MizBuilders we're better off coming up with 90 min battles which when complete LOAD another 90 min battle mission and so on. Its still possible to create immersion & reality on smaller scales its just how you go about it.

[Speed]

I have a CAS mission which will take you over 4 hours to complete if everything goes well. I must admit with 180 triggers its starting to cause a few FPS problems so I might have to hack it into pieces.

 

I have however through scripting managed to add a LOAD & SAVE option so you dont have to restart the whole mission from the begining each time. Now it not Dynamic as previously destroyed units won't appear again in their destroyed position smoking away but its not really required for this type of CAS mission as I have you moving around the map supporting ground units.

 

I understand Wildfires passion for creating a large scale battlefield environment but to be honest when tasked you would only be dealing with a small part of it IRL. Having multiple units fight it out even 80nm away from your local battle is only going to use up unecessary CPU utilisation.

 

DCS just isn't designed to handle large scale battles so for now as MizBuilders we're better off coming up with 90 min battles which when complete LOAD another 90 min battle mission and so on. Its still possible to create immersion & reality on smaller scales its just how you go about it.

 

I've had well over 1000 triggers in a mission and had no FPS hit from them. While it depends on the exact nature of the triggers, my personal estimate would be you would need at least 5000 triggers before you took a noticeable FPS hit from them. 180 triggers ain't nothin to worry about- unless they have like... 30+ conditions on them. I once tested a script that went through like 1,000,000 calls per second to the game engine to get terrain altitude at various points along a line. It only caused like a half second pause. I don't think 180 triggers is gonna phase the game engine much :thumbup: It's things like active AI and calculating bullet trajectories that slows down the game most.

 

As far as what you say about the 90 minute missions... I kind of agree.

 

Personally though, a dynamic tasking engine of some sort is on the list to be created... first, I gotta finish and release this new mod I'm working on, then a video guide for that mod, then a scripting guide for all five Lua environments, then an update to that mod that allows users to call in airstrikes. Once that's done, then I can start on this dynamic tasking system I want to make. The dynamic tasking system will assign tasks to players based on actual battlefield conditions. So if, like, a group of M1s spot some enemy tanks, they will try to halt outside of the enemy tank's firing range and call in air support before they continue. If they come under fire by enemy tanks, then they will try to call in CAS. Predators circling overhead will be able to automatically call you in on targets they see, based off of probability of detection and line of sight calculations. That kind of stuff. Should lead to big battlefield missions with a huge amount of variability.

Edited July 11, 2011 by Speed

[WildFire]

Yeah like I said long or big missions are easy to create, you can get carried away in the editor in a matter of minutes... I know what druid is saying though and its understandable, I dont necessarily want the whole country in turmoil, that of course would be insanity, Im saying more like an actual FLOT and perhaps a war going on around that. Say you are within 30 miles of a certain area there is an active battle going on.

 

Going to have to find a way to keep it going with scripts though, such as groupA is engaged with groupB, when either one is groupdead then spawn new group and move them into the battle. Do so with both sides. So the battle never really ends and is still somewhat localized around a central point. Perhaps you could have 3-5 medium sized groups of 50-100 units spread across 20 klicks then you effectively have a FLOT. It wouldnt cover the whole map by any means but at least it would be a much larger and more immersive environment than we have yet to see.

 

The only problem with doing this now is that smoke burns for 20 minutes on your standard tank/apc so you would effectively kill the GPU with all the smoke being rendered(cause tanks are pretty excellent at killing each other depending on engagement distance and skill level) plus like speed said the bullet calculations and instructions for 300-600 active units would render the engine relatively useless... maybe... I'll have to try that out and see what happens.

 

So far in ShatteredGlass there is a task that has 60+ pieces assaulting an airbase with about 20 pieces + infantry defending and when the tank battalions start taking a bit of losses and the battle is hot and heavy that is when people are reporting spikes. Maybe just too much info to send across a standard network. Hence trying to limit what I can to streamline the whole package. This mission has about the least amount of triggers as can be made and im kind of surprised it all works. So background processes shouldnt be eating up a whole lot here.

 

Anyhow good input, excellent point of views, and apparently speed should be on the payroll. :thumbup:

[snwboardn]

 

Personally though, a dynamic tasking engine of some sort is on the list to be created... first, I gotta finish and release this new mod I'm working on, then a video guide for that mod, then a scripting guide for all five Lua environments, then an update to that mod that allows users to call in airstrikes. Once that's done, then I can start on this dynamic tasking system I want to make. The dynamic tasking system will assign tasks to players based on actual battlefield conditions. So if, like, a group of M1s spot some enemy tanks, they will try to halt outside of the enemy tank's firing range and call in air support before they continue. If they come under fire by enemy tanks, then they will try to call in CAS. Predators circling overhead will be able to automatically call you in on targets they see, based off of probability of detection and line of sight calculations. That kind of stuff. Should lead to big battlefield missions with a huge amount of variability.

 

YES! That is exactly what I am talking about! You need to get on DCS's payroll with all that work you are putting in. ;)

 

Of course with that you need a partner to help you with a sound mod, make it sound a little more hectic on the ground with gunfire in the background. :D

 

edit* that is funny I didn't even see your post wildfire when I wrote mine

Edited July 11, 2011 by snwboardn

[Speed]

YES! That is exactly what I am talking about! You need to get on DCS's payroll with all that work you are putting in. ;)

 

Of course with that you need a partner to help you with a sound mod, make it sound a little more hectic on the ground with gunfire in the background. :D

 

edit* that is funny I didn't even see your post wildfire when I wrote mine

 

 

Uh, I'm not anywhere near the time commitment or expertise level that people who actually get paid to write code are. I appreciate the compliment, but it's not really true. I also still have a lot of work to do... only thing I have "officially released" was chatIOlibv1.

[Ripcord]

Wildfire, the answer to your question probably lies in the settings of your clients on the server that are flying your mission. Are they flying with labels on? Or maybe they should set the scenes to medium, if they are experience slowdowns online? The problem may not even be related to number of units or the content of the mission itself.

 

That said, I have heard that spawing objects/groups does create slowdowns while in the heat of battle, even in SP -- are you spawning all your groups at mission start, or pretty fairly on?

 

I agree with Druid, I don't believe it is just a function of lingering smoke -- anyway that is all part and parcel of the battlefield environment you are trying to create.

 

This thread is interesting for me because I am creating a large FLOT type of mission, two really, in which side A is going the offensive and side B is defending a position, trying to prevent the enemy of reaching their objective. My version is maybe not as big as yours, say 200-250 units on each side, and not all units are directly involved right at the point of attack -- close to half are spread out across a 30-40 km FLOT. Hell, some of them are just AirDef units, or FARPs not even in the area. Trying to create a good theatre template that works well. Then I can just move my units around, change it up and reuse. It may one day morph into a nice campaign, but we will see. Right now I am kinda like you, just trying to get one mission working well.

 

How many players in MP do you have in your mission? Is it 2 players or 4 players or more? Does it make a difference in frame rates if less players fly it? Just throwing out ideas.

 

Ripcord

[Druid_]

I've had well over 1000 triggers in a mission and had no FPS hit from them. While it depends on the exact nature of the triggers, my personal estimate would be you would need at least 5000 triggers before you took a noticeable FPS hit from them. 180 triggers ain't nothin to worry about- unless they have like... 30+ conditions on them. I once tested a script that went through like 1,000,000 calls per second to the game engine to get terrain altitude at various points along a line. It only caused like a half second pause. I don't think 180 triggers is gonna phase the game engine much :thumbup: It's things like active AI and calculating bullet trajectories that slows down the game most.

 

Interesting. Maybe there is something else in there causing problems. At one point during one task there is a single 3-5 sec freeze & I am almost at a loss as to what is causing it since it doesn't coincide with any triggers or anything else I can think of. It only happens during this one task. I will have to test further.