We have a MonoBehaviour component holding a long list of prefabs (can be 100).
Those prefabs are meant to be created/destroyed as needed.
The component is singleton getting created on first usage.
The question is related to the memory usage, using a mono behaviour component, I understand this means all the prefabs are being loaded onto the disk when creating the component.
I am looking at a way to limit the loading of those rarely used prefabs.
I thought of having them in different scriptable objects so only the lists from used prefabs would get loaded, with the lists of less common prefabs most likely rarely or never loaded.
I am looking at tips on the topic.
Related
Whenever I need to instantiate prefabs, I always use the following technique
Add a GameObject/MonoBehaviour field to the MonoBehaviour which would instantiate the prefab.
Use this field to instantiate the prefab
E.g.:
public class ExmapleController : Monobehaviour
{
[Serializefield]
GameObject _examplePrefab;
GameObject _exampleObject;
public void SpawnExample()
{
_exampleObject = Instantiate(_examplePrefab);
}
}
Now my problem with this is that first of all I need two references. One for the prefab and one for the instantiated object. Secondly, whenever I need additional prefabs I have to create new and new references inside my script. This sometimes clutters my Scripts and I always wonder that is this memory efficient? What does Unity do behind the scenes? It just only stores a GUID reference for my prefab than somehow loads this prefab based on this reference? Or when I make a prefab reference the whole object tree is pulled into memory and I should rather use Asset bundles or another Unity technique?
Basically that's fine. But if you want to improve your workflow you can try creating ScriptableObjects to store the prefab references and then have only 1 reference to your scriptable object inside the mono behaviour. It will help you keep your mono behaviour classes clear. For more advanced approach you can check Dependency Injection frameworks for Unity. The most common one is Zenject. It helps you in managing dependencies in your classes (prefabs in your case are dependencies of Monobehaviour class).
When you create a reference to a Prefab Unity only store the GUID of the prefab you referenced. It is how resource system in Unity works. But if your prefab is required on the current scene then it will be fully loaded with all inner dependencies. If loading takes much time than it is the reason to think about using Addressables (AssetBundles). Unity doesnt have any other technique to avoid loading the full object tree in memory. But addressables let you control when to load the specific prefab from asset bundles to memory.
Yes this perfectly fine and how it's done. A reference is just that: a reference. The only time a copy of the object is made is when you call Instantiate.
If you have multiple prefabs that are just variations of one you should use ScriptableObject instead and have just 1 prefab that can take any variation form. You would still intantiate them the same way.
Basically that's fine. But if you want to create a huge numbe of objects or delete them you could also have a look at Object Pooling. More infos can be found under the link from the tutorial at Unity:
https://learn.unity.com/tutorial/introduction-to-object-pooling#
"Object Pooling is a great way to optimize your projects and lower the
burden that is placed on the CPU when having to rapidly create and
destroy GameObjects. It is a good practice and design pattern to keep
in mind to help relieve the processing power of the CPU to handle more
important tasks and not become inundated by repetitive create and
destroy calls. (Unity)"
I have four different Canvas elements in my Unity scene. This organizes them by concern, but also improves performance since I don't have to rebuild all the elements if I "dirty" one of them. I want to create prefabs of each so I can reuse them across the game, and also so there are less elements in my scene file (I have yet to figure out why Unity doesn't re-organize YAML to match the scene order and hierarchy. Currently I can see no reason and it makes source control a huge pain).
So, my question: Should I include the Canvas as the parent GameObject in the prefab, or should I make the first child the prefab (see image)? I see two perspectives:
On one hand it's nice to scan the list of elements in the root of the scene and see the blue objects and know they are automatically replicated across scenes. Plus, the more object stored in the prefab then there will be less to pollute my YAML file.
On the other, with a Canvas Scalar script on the parent, the sizes inside the prefab are going to be updated every time I change the resolution / platform and I've read it's not good always update them outside of prefab mode. The other issue is the fact that UI elements usually have a lot of external references to update text and so forth.
After a lot more research I've come to some conclusions about the answer:
As stated in this talk at Unite 2017, isolating UI element "islands" by concern AND how often they are updated will keep the UI from affecting performance. The video also states it is fine to use multiple Canvases inside other Canvases to achieve this goal.
If you add a prefab whose values are updated from outside the prefab they will appear as "overrides" to that prefab. This includes if you link to elements outside the prefab from within it and vice versa (see image). Importantly, it is not possible to apply these overrides to the prefab, and that's kind of the point. Unity accepts that these changes will happen and this is how it deals with them without breaking the prefab or the scene.
So the answer is yes, it's no problem to include the Canvas parent in the prefab, and you can even organize it inside other Canvas so they all inherit the same settings. Further, since the more code stored in prefabs decreases the extra (unorganized) identical code in the .unity YAML file (in my case from 17,000 lines to just less than 5000!) then this is the recommended way.
Firstly, I'm new to Unity and I'm currently learning c#, please be gentle!
As a project, I'd like to create a simple 3d platform game. The idea being the player starts on a square tile, which is the game platform playing area. They explore different elements on this tile by moving around.
The perspective is 3rd person, so the player is looking down on the action. As they approach the edge of the tile Unity recognises this and adds another tile to the existing one, basically doubling the playing area.
As the player moves around further to the tiles edge, other tiles are added, increasing the discovered areas each time.
My thinking to achieve this would be to have 5 tile game objects, that have different elements like trees already built on them. The question is what is the best way to achieve this.
Would it be to build the complete level with all tiles and then using code restrict how far ahead the player can see, basically to the width and length of the tiles.
Or would it be better to trigger a new tile to appear as the player approaches the edge of the current tile.
Thought, links to example code that I could recycle would be very handy.
Thank you.
The question is primarily opinion based, but here is some advise:
There is no right way to make a game in unity. The correctness is mainly focused on performance; memory and processing.
There are two approaches here:
condition: Your world is endless (or very large) and player moves very slowly over time (say each tile takes more than a second to traverse).
solution: Do not instantiate everything at start, instead when player reaches the edge of a tile look for the next tile in your level structure to retrieve its data and then instantiate its objects.
condition: Your world is not very large (say it has less than overall 10K objects) OR player can move fast (as in Age of Empires game).
solution: Instantiate everything in loading phase and deactivate the game objects as they are being instantiated so that nothing is getting processed at the start of the game. In this case the level structure is a collection of game objects where you activate or deactivate them.
Hints:
Pooling is a good practice when it comes to repetitive objects. Pool is basically an empty game object with a script, and many many deactivated children.
e.g. If you have 1K trees of the same kind you better have a tree pool. Create a game object named tree pool, add a pool script to it and make that script generate 1K deactivated tree game objects in the loading phase. Then whenever your level generator needs a new tree just fetch a tree from tree pool and activate and reposition it. Make sure you clean up the pool after the game ends to prevent memory leak.
you can implement the pool as you like to provide object variations for a specific pool.
Instantiation in run-time is costly especially in mobile devices. Whenever you need to instantiate several game objects at once consider using a coroutine to prevent lag or freeze (this applies to loading phase too).
e.g.
IEnumerable CreateObjects(Data[] data)
{
foreach(var e in data)
{
Instantiate(e);//instantiate here and then wait
yield return null;//this line prevents lag
}
}
//...
StartCoroutine(CreateObjects(data));
//...
Having many active game objects is also costly especially in mobile devices. If many active game object have heavy scripts, the update methods will ruin performance. If many active game objects have rigidbodies or colliders the physics engine processing gets heavy.
Activating game objects is costly but less than instantiation, since it only runs Start method on its script if any, and also causes the physics engine to re-evaluate its structure if it has static collider or rigidbody.
Rendering is costly if only the rendered visuals are (fully or partially) visible to the cameras.
Create an editor project where you can design your levels. in this project you will make use of unity's editor classes to generate some kind of data (JsonObject) which represents your whole world. (collection of levels, tiles in each level and objects in each tile). Moreover you can have other data such as objectives and stuff each stored in a different JsonObject. Then use this data in your game by reading it and storing its structure in memory.
The reason I recommend JsonObject to store data is its size, portability, flexibility, accessibility and simplicity to work with. JsonObject is a dictionary stored as a plain text. and since it's a text you can encrypt it as you like
I have some confusion with scenes in Unity3d and I was not able to find any resources about them.
When should scenes be used? For example in a platformer would every level have to be a different scene? Would the main menu be a scene?
Can one overlay scenes?
How do assets work between scenes? Are they attached to each individual scene and have to be reloaded every time. Can one specify when an asset is no longer needed?
How does one send data between scenes/interface between scenes?
I understand that this is a broad topic, but I didn't want to spam with multiple questions.
When should scenes be used? For example in a platformer would every
level have to be a different scene? Would the main menu be a scene?
There are no general rules about that. In theory you may have just one scene for the whole game.
How you organize your scenes is entirely up to you and often depends on the type of game you are creating.
I think that there are at least 3 features to be considered of using scenes:
they are a logical container for all pre-instantiated objects that might be useful to divide your game into multiple levels/sections.
You can serialize cross references between GameObjects and Components inside a scene (if GO A needs a ref to GO B, and they belong to the same scene, the reference can be serialized and you no longer need to find the referenced object at runtime)
When you load (not in an additive way) another scene, the resources already loaded into memory are automatically released
Can one overlay scenes?
Yes you can using LoadAdditive. Unfortunately, once 2 scenes are both loaded into memory there is no automatic way of distinguish objects belonging to one or the other. So if you load additive a second level environment, it's up to you to keep track of the previous environment and explicitly destroy it if you need to.
How do assets work between scenes? Are they attached to each
individual scene and have to be reloaded every time. Can one specify
when an asset is no longer needed?
As defaults every GameObject of a scene will be destroyed once the new scene is loaded (unless you use an additive scene loading). A way to make a GameObject survive across scenes is to mark it using DontDestroyOnLoad.
If you need to share a particular "configuration" of a GameObject, you can store it as a prefab, and reference it across scenes (but remember that once in a scene it's a prefab instance, so the GO shares with the prefab the initial serialized and not overriden properties, but 2 instances of the same prefab are different objects).
How does one send data between scenes/interface between scenes?
Several ways, depending on what kind of persistent data you want to share.
For a particular GameObject instance let the object survive using DontDestroyOnLoad.
If you have some configuration data that doesn't need to be attached to a specific GameObject you can consider storing a ScriptableObject inside the AssetDatabase and reference it.
If you have data that must persist across different game sessions you can consider storing them into PlayerPrefs.
There are 2 other ways that I don't like, but just to cite them:
Using a static field can sometimes help you in doing that, but it has several problems from my point of view
Save and load from disk (could be useful in several situations, but often it's a platform dependent way and you can have some trouble especially on different mobile platforms)
This is a broad topic btw, I hope this answer can be a quite decent overview.
When should scenes be used? For example in a platformer would every level have to be a different scene? Would the main menu be a scene?
There is no rule as to how many scenes you need to have in your game. However, scenes allow you to logically separate out parts of your game from the rest of it. You have to have a minimum of one scene.
By main menu, if you are referring to a canvas with your UI elements, it will be IN a scene and not a scene itself. Canvas is just another GameObject, that we mostly happen to use for showing game menus. I mostly create a Canvas GameObject, put a script by the name of "UIManager" and put DontDestroyOnLoad on it, so I have access to it in all scenes. Make it Singleton and I ensure that it is not duplicated.
Can one overlay scenes?
Yes, there is no restriction as to how many scenes you can load at a time. What purpose do you plan to overlay scenes though? Maybe there is a better way than loading additively.
How do assets work between scenes? Are they attached to each individual scene and have to be reloaded every time. Can one specify when an asset is no longer needed?
Assets are what you see in your 'project' hierarchy. I think you meant "GameObject"s in the scene, and if so, think of your gameobjects as entities with components (Entity-Component System). All entities in a scene get destroyed when its parent scene is destroyed until explicitly stated not to, using DontDestroyOnLoad in some component (a monobehavior in case of unity). The destroyed ones will get garbage collected.
So how they are loaded (or reloaded) depends on your implementation, on whether you are instantiating/destroying them time an again or if you put their instantiated prefabs in a cached object and retrieving later from it.
How does one send data between scenes/interface between scenes?
Heisen covered the ones I could think of. Just to add a little bit to it, it also depends on how you want to the architect your project. So if you had an underlying data structure to e.g. hold Commands, you are free to use it in any part of your project
Most games would be organised to have scenes for every level(including the main menu) but that is entirely up to you.
You can use the data from one scene to another if you save it in a text file or binary. There are a lot of tutorials on how to do this. I find documentation helps a lot.
Assets are universal in a project.
You can not overlay scenes.
When should scenes be used? For example in a platformer would every level have to be a different scene? Would the main menu be a scene?
When to use a scene is up to you. If you are just starting I would recommend using a different scene for each section of your game.
Can one overlay scenes?
Yes, using LoadSceneMode.Additive(). (LoadAdditive() is obsolete)
How do assets work between scenes? Are they attached to each individual scene and have to be reloaded every time. Can one specify when an asset is no longer needed?
By default, assets are deleted when using SceneManager.LoadScene(). However, if you use DontDestroyOnLoad(), the object will not be destroyed when entering new scenes. If you want to only keep an object through a few scenes instead of all, use Destroy() with some boolean logic.
How does one send data between scenes/interface between scenes? I understand that this is a broad topic, but I didn't want to spam with multiple questions.
You can send data through scenes by using the aforementioned DontDestroyOnLoad(), referencing the data on different scripts, using ScriptableObjects, using JSON Serialization, using StreamWriter(), using PlayerPrefs (Don't use for important information), the list goes on. I would personally recommend using ScriptableObjects for their accessibility, and StreamWriter() for it's encryption capabilities.
In my game I have thousands of "tile" nodes which make up a game map (think simcity), I am wondering what the most frame-rate/memory efficient route for texturing and animating each node would be? There a a handful of unique tile "types" which each have their own texture atlas / animations, so making sure textures are being reused when possible is key.
All my tile nodes are children of a single map node, should the map node handle recognising a tile type and loading the necessary atlas & animations (e.g. by loading texture & atlas names from a plist?)
Alternatively, each tile type is a certain subclass. Would it be better for each SKSpriteNode tile to handle their own sprite atlas loading e.g. [tileInstance texturise]; (how does sprite kit handle this? would this method result in the same texture atlas being loaded into memory for each instance of a certain tile type?)
I have been scrounging the docs for a deeper explanation of atlases and texture reusage but I don't know what the typical procedure is for a scenario like this. Any help would be appreciated, thanks.
Memory first: there won't be any noticeable difference. You have to load the tile's textures, textures will account for at least 99% of the memory of the Map+Tiles and that's that.
Texture reuse: textures are being reused/cached automatically. Two sprites using the same texture will reference the same texture rather than each having its own copy of the texture.
Framerate/Batching: this is all about batching properly. Sprite Kit approaches batching children of a node by rendering them in the order they are added to the children array. As long as the next child node uses the same texture as the previous one, they will all be batched into one draw call. Possibly the worst thing you could do is to add a sprite, a label, a sprite, a label and so on. You'll want to add as many sprites using the same texture in consecutive order as is possible.
Atlas Usage: here's where you can win the most. Commonly developers try to categorize their atlases, which is the wrong way to go about it. Instead of creating one atlas per tile (and its animations), you'll want to create as few texture atlases as possible, each containing as many tiles as possible. On all iOS 7 devices a texture atlas can be 2048x2048 and with the exception of iPhone 4 and iPad 1 all other devices can use textures with up to 4096x4096 pixels.
There are exceptions to this rule, say if you have such a large amount of textures that you can't possibly load them all at once into memory on all devices. In that case use your best judgement to find a good compromise on memory usage vs batching efficiency. For example one solution might be to create one or two texture atlases per each unique scene or rather "scenery" even if that means duplicating some tiles in other texture atlases for another scene. If you have tiles that almost always appear in any scenery it would make sense to put those in a "shared" atlas.
As for subclassing tiles, I'm a strong proponent to avoid subclassing node classes. Especially if the main reason to subclass them is to merely change which texture they are using/animating. A sprite already is a container of a texture, so you can as well change the sprite texture and animate it from the outside.
To add data or additional code to a node you can peruse its userData property by creating your own NSMutableDictionary and adding any object you need to it. A typical component-based approach would go like this:
SKSpriteNode* sprite = [SKSpriteNode spriteWithWhatever..];
[self addChild:sprite];
// create the controller object
sprite.userData = [NSMutableDictionary dictionary];
MyTileController* controller = [MyTileController controllerWithSprite:sprite];
[sprite.userData setObject: forKey:#"controller"];
This controller object then performs any custom code needed for your tiles. It could be animating the tile and whatever else. The only important bit is to make the reference to the owning node (here: sprite) a weak reference:
#interface MySpriteController
#property (weak) sprite; // weak is important to avoid retain cycle!
#end
Because the sprite retains the dictionary. The dictionary retains the controller. If the controller would retain the sprite, the sprite couldn't deallocate because there would still be a retaining reference to it - hence it will continue to retain the dictionary which retains the controller which retains the sprite.
The advantages of using a component-based approach (also favored by and implemented in Kobold Kit):
If properly engineered, works with any or multiple nodes. If what if some day you want a label, effect, shape node tile?
You don't need a subclass for every tile. Some tiles may be simple static sprites. So use simple static SKSpriteNode for those.
It lets you start/stop or add/remove individual aspects as needed. Even on tiles you didn't initially expect to have or need a certain aspect.
Components allow you to build a repertoire of functionality you're going to need often and possibly even in other projects.
Components make for better architecture. A classical OOP design mistake is to have Player and Enemy classes, then realize both need to be able to shoot arrows and equip armor. So you move the code to the root GameObject class, making the code available to all subclasses. With components you simply have an equipment and a shooting component add to those objects that need it.
The great benefit of component-based design is that you start developing individual aspects separately from other things, so they can be reused and added as needed. You'll almost naturally write better code because you approach things with a different mindset.
And from my own experience, once you modularize a game into components you get far fewer bugs and they're easier to solve because you don't have to look at or consider other component's code - unless used by a component but even then when one component triggers another you have a clear boundary, ie is the passed value still correct when the other component takes over? If not, the bug must be in the first component.
This is a good introduction on component-based design. The hybrid approach is certainly the way to go. Here are more resources on component based design but I strongly advice against straying from the path and looking into FRP as the "accepted answer's author" suggests - FRP is an interesting concept but has no real world application (yet) in game development.