Precaching resources and load times - performance

I've been doing some research on caching resources for a game environment. By analyzing various games both 2D and 3D, I've come up with some questions.
In 2D games, such as Link to the Past, we see seemless transition from overworld to dungeon, or from house to overworld etc. I would imagine this means that all the game's resources are precached. That's all fine for 2D games since their resources tend to be smaller, but what about 3D games?
Half-Life 2 and Assassin's Creed for example, both have a loading screen inbetween map transitions. I know HL2 uses a sound cache, where a 5 second clip of the start of music will play while the actual mp3 loads. But this doesn't answer questions about textures and models. Do these games precache these resources to speed up load times, or do they not to save memory?
Further, we see 3D games like Twilight Princess where there is again seemless transition. Is this implying that all resources are cached here as well? I can understand this a bit more for PC games, but on the Wii of all platforms?
Overall, what is the best option to use for game development in both 2D and 3D environments?

Loading screens are what happens when a game's resources are moved from being on disk to being in RAM, while something like Twilight Princess uses something akin to Lazy Loading, where parts of the world are dynamically loaded into the game (while unloading other parts of the world to keep RAM usage low). Resources aren't necessarily cached, so much as managed differently.
In the case of a 2d game like Link to the Past (or Super Mario), there ARE loading screens, but they're short enough and often hidden by Link entering a dungeon, or the level/amount of lives left message in Mario, or the animation of going through a door in Metroid. Also, and perhaps more accurately, cartridge based games are effectively solid state memory, and the memory access speed is equivalent to RAM.
As always, the best option depends on what you're making. Ideally, you want to prevent ANY loading screens, but that's usually hard/impossible, especially with a large 3d game with tons of assets. While some games are clearly divided into levels (Portal especially comes to mind here), an open world game wouldn't be any good if you had to sit through a loading screen every city block. 2d games are a bit easier, since you don't have to deal with complex model data, but many of them have some sort of loading screen, usually when you start the game. The more complex the game, the more likely you'll have to deal with some sort of loading screen.
For great examples of 'lazy loading' in games, check out (the 3d) Grand Theft Auto games, Minecraft (infinite worlds), and Ultima 7 (the massive gameworld was divided into regions which got loaded when you go close to them)

Related

Unity - How to get started working on a large terrain/map?

I want to begin working on a big sandbox game with lots of oceans and islands. This will obviously require a very big map.
I've spent the last couple days researching the best methods on doing this and so far have figured out that
I need to split the terrain into tiles/chunks.
I need to have the player only load those tiles/chunks that are closest to them and unload those further away.
That being said. I would rather avoid flooding my game with a bunch of assets and want to kind of handle things on my own and hard code these systems in myself.
Some questions,
Can I just create a massive 20k x 20k terrain using Unity's built in terrain editing system and then worry about splicing and loading/unloading the tiles later?
or do I need to build my big terrain in an alternative program and import my terrain in then handle the splicing and loading situation?
Also, when it comes to multiplayer. I assume I would just need to basically do the same thing just for each client?
I would appreciate any other tips or guidance on doing this as well. Thanks.
Splitting into multiple chunks but ones that arent too small seems to be the better option, maybe like 4 quarters, that way you are not flooding with assets, you would disable all quarters except the one the player is in.
The problem with making a big terrain is, while you could get a tool from the asset store to split ur terrain later, you will get into the headache of resolutions and sizes and overall I guess its better to split it into 4 pieces or a bit.
You would have a trigger point where you want the game to load the terrain, and depending on which trigger point the player entered you would load the corresponding terrain piece.
For multiplayer, you would do the same thing for each client, making your big terrain as 4 smaller terrains and then loading the corresponding quarter. Tell me if I missed something.

DirectX 11.2 in Visual Studio 15 [closed]

Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 5 years ago.
Improve this question
My friends and I go to a vocational school for computer programming. We got together and decided to start making games, but the problem is only two of us have. I've made a text adventure game using lots of buttons and lots of forms in Visual Studio, and my other friend has made a few games using GameMaker. All of us have the same basic knowledge of Visual Basic (due to being in the same Programming course) and we want to try our hands at making a 3D game. I've done a little bit of research, and determined that we may need to use DirectX if we wish to keep using Visual Studio.
The problem is, I don't know how to start really. I've tried looking it up. Is it possible to access the API using just Windows and Visual Studio?
Also, based on what I can tell, we need to know C++. If that's true, then we'll convert to C++ to make games. It shouldn't be too difficult, I'd imagine.
Planning on making a game in 3D? This is no easy task and is very daunting. This isn't to scare you away from pursuing it, but to let you know what is down the road ahead of you. I'll break this down into several lists to give you a clear or better understanding.
Things to Consider
What type of game are we going to make?
2D Side Scroll, 2D Bitmaps and Sprites, Partial 3D or 2D Overhead View?
If game is 3D as you have stated; what is the view style: First Person, 3rd Person, View and Zoom, Overhead, Moveable Camera with auto follow or fixed?
Is it an adventure game, a role playing game, first person shooter, strategy?
What are the rules of the game?
What are the challenges, achievements, goals, and scoring or rating systems?
Is it single player, multi player, or mass multi online player?
Does it have levels or boards, or is it a sandbox open world?
Is there a story, plot and main characters?
What is the ultimate achievement or the final goal?
Does the game infinitely progress?
What type of background music and sound effects should it have, such as fantasy fanfare, drum beats and trumpets such as a war march, mystical, suspense and horror?
Who and what is our target audience?
Is this game for PC, Console, Tablet, Smartphone, Web browser?
Is it for Windows, Apple-Mac, Linux-Unix, Android Devices or all of them or just some but not others?
This deals with choosing the API to use either it be DirectX, OpenGL, or others -
DirectX - Which version? 9.0c, 10, 11, 12? This will affect target audience since different version work with different operating systems of windows such as DirectX 12 only works with Windows 10. Also not everyone has a DirectX 12 capable video card.
OpenGL - The version only makes a difference to some extent since almost all hardware vendors must support at least the minimal Core API library, yet they are each allowed to support their own independent list of supported extensions.
Others - Such as the newly formed Vulkan, the same concepts from above can be applied.
Is the target audience for younger children say between 3-7? Is it for children say 5-12? For teens 12 - 17? Mature for adults 17 or 18+?
Is the game geared to gender specific? Such as is it a game intended for boys or men such as wrestling, boxing etc., or is it more intended for girls and women such as Barbie Dress Up or Horse Grooming? This isn't meant to be stereotypical, just something to consider when designing a project.
Choosing an API
If you know your target audience is primarily using Windows, then 9/10 times DirectX is the choice to go, albeit Modern OpenGL works just as good!
If you know that you are targeting window users, but also want to include non-windows systems, then you may want to use OpenGL since DirectX isn't available for other operating systems. The best bet would be to use both API's and have a rendering switch based on the OS in use, or give the user the option during Game Installation if PC to choose which version of the Game such as DirectX or OpenGL. If it is Windows Phone or iPhone or Android, then OpenCL is probably your bet then.
Choosing a programming language
This all depends on the particular needs.
The most commonly used language for large scale 3D Game Engine projects
is primarily found using C++ because of the language's versatility in allowing you to do many different tasks not available in other languages such as managing your own memory dynamically, writing generic programming through templates, object oriented by using classes, inheritance and polymorphism or using a base classes to design abstract class at the root of a class hierarchy.
C# is not without its limits for it does allow for easier code writing without having to worry about manually managing memory and other things.
Java is still widely used as it's virtual machine is found on almost every device.
JavaScript is excellent when working with any major browser such as IE11, Chrome, Firefox etc.
Visual Basic is one of the least used, but it still has it's merits:
For example: There does exists a program to make 2D RPGs out there and it is called RPG Maker. With the use of text boxes, lists, radio buttons etc. VB would be a great fit to design the application part that makes the games, however it also has it's own scripting language that you can program game content from its own scripting language. In many cases these game scripting languages can be written from almost any language such as C, C++, C#, Ruby, Pascal, Pearl, etc.
Choosing appropriate 3rd Party Libraries that are well known, stable, reliable, trustworthy, constantly updated, flexible and portable. One such as GLM which is an open source math library that was written with the feel and style of OpenGL's GLSL shading language. It is constantly updated, works very well, is fully portable, easy installation as it is a headers only library; no need for other dependencies, binaries or having to recompile it for each different version of the compiler you are using etc.
The Assets
What kind of theme does or will the game have?
The Art - Graphics, Textures, Models, etc. - The overall look and feel.
The Audio - Music background, sound effects etc. - The Music Compositions.
Storyline: - Cut Scenes, Video, Animation, Character Dialogs etc.
Game Control and Logic
What kind of controls? - Keyboard & Mouse, Joystick, Hand Held Controller?
The actual motion of the Characters or Game Objects. Kinematic motion, point and click, point and drag, hover over, etc.
The Actions of the Characters or Game Pieces.
The logic of the game: If this then that etc. Scoring System or Goal Achievement
Math-Math & More Math - Without it; Good Luck!
Basic Algebra
Geometry
Trigonometry
Logarithms, Exponentials, Summations and Matrices.
Points, Vectors and Matrices
Calculus - Differentiation: Derivatives, Partial Derivatives; Integration: Integrals, Integrands, both definite and indefinite integrals.
Linear Algebra
Vector Calculus
Vector Fields
Abstract Algebra
Analytical Geometry (Geometrical Projections) - Non Euclidean.
Physics - Kinematics, Quadratics, Forces, Vector Fields, Collisions, Explosions. Just for game content. Electrical Circuitry for circuit boards: need to know your systems and how ROM, RAM, CPU, CACHE, VRAM, and GPUS all communicate across the BUS. Without this you can not "Optimize" your Game Engine!
The Tools Needed
A high powered pc workstation that is capable of doing graphic, video and audio manipulation with ease.
The appropriate OS with extended cards: Video, Audio, Networking etc.
The programming languages and their compilers. C++, C#, Java etc. Any well known IDE.
A graphics editor such as Photoshop, Gimp, Texture Editor etc.
Modeling programs such as Maya, 3DS Max etc.
Audio as there are two varieties:
Orchestra composition for making music scores to export into 3D Audio files.
Sound Effects to output audio types such as *.wav files.
All of the appropriate 3rd party libraries to simplify the amount of code you will have to write.
A Hex Editor for reading and manipulating Binary Files!
A Programming Calculator - For doing both binary and hexadecimal calculations, words, and bytes etc.
A versatile 3D graphing calculator for doing advanced mathematics. You may need to solve a system of equations with multiple variables to find either their derivatives, 2nd order derivatives, partial derivatives, integrals, doing trigonometric equations, vectors, and matrices. Especially Vectors and Matrices using trigonometry for doing rotations, and more advanced concepts such as quaternions and maybe even complex numbers although I have rarely seen the need for complex numbers in any Game Engine, except maybe for some concepts such as in particle physics and AI-programming.
The Concepts of a Game Engine
The Parts - In todays industry a Game Engine isn't just one thing, it can be considered multiple little engines integrated together into one working engine.
Graphics Engine - Window Creation, and API setup for rendering Primitives to the screen each game frame while considering the monitors' refresh rate and its pixel resolution as well as the GPUs draw rate.
Textures - Normally a 2D graphic images stored in memory with a width and height either measured in inches or pixels, and can have various properties or attributes such as color encoding, and compression levels.
Sprites & Fonts - Are usually associated with a texture since they are in a sense a specific type of texture.
Models - Models are usually 3D Geometric Meshes composed of mostly triangles, sometimes lines, points and other primitives where a specified texture can be applied to them.
Particles - Particles and even some aspects of animation can be applied here although the particle engines along with animations are more associated with the physics engine because they apply physics from the physics engine.
3D Environment - Could be in a closed room, outdoors, or outer space or regular flight.
Camera - The view angle at which the User will see the Game World From.
First Person - The illusion as if you were looking through the main characters eyes. The main character and camera turn and move together in unison.
Third Person - The main character is fully in front of the camera. The camera will rotate around the point of origin with that origin being the main character unless independent controls are given, the camera will follow or translate linearly as the character moves. It may also have a zoom option. Also the tilted view is how much angel does the camera project between the main characters direct line of sight (horizontal) versus how much the camera is pointed towards the ground.
Materials - The color attributes of a model's or a triangle's face: It is usually defined by 4 color channels (RGBA) where RGB are the Red Green and Blue channels and A is the alpha or the amount of transparency it has. Materials are important because they affect the overall shade of the object. If you have a model of a box loaded into memory that has an all yellow face and you apply a blue material to it, through color addition and subtraction it would now appear green. Materials are also useful and important when it comes to lighting which you will see in the next section which are described by these kind of material - lighting combinations: Ambient, Diffuse and Specular colors or material properties.
Ambient - The color and intensity that the light puts off.
Diffuse - The color saturation of the object itself (Self emitting)
Specular - The shininess of a region of the object that changes color from the light intensity on shiny type objects such as an apple or metal.
Lighting - Much can be said about this but without it you would not be able to see an interesting setting and you would have minimal visual effects.
Point Light - Has location but no direction as it shines out from a single point such as a light bulb.
Directional Light - Has direction but no location such as the sun
Spot Light - Has both direction and position as well as a cutoff angle. Such a light is like a flash light. It also attenuates.
Properties - Some lights attenuate, others have intensities, specular highlights, cutoff angles, etc.
Shadow, Reflections etc. - This is the ability to use the above lights
materials, their colors, directions, and intensities etc. to cast either a shadow of an object or even to reflect objects such as the sky or clouds onto a surface of shiny metal or a body of water.
Weather Patterns - Fog, Rain, Snow, Fires etc. Now Fog is usually done with shaders, and so are the others, but in most cases things such as rain, snow and fire are usually generated with a particle engine.
Motion - Transformations - Some of these are tied into the game logic of the controls to be used, some are tided into the scene graph such that if 1 object has multiple parts and the whole object is translated all the parts are translated together, and some of these are tided in with the physics engine for things such as collision detections and even the AI system for enemies to check for boundaries and to be aware of the player as being hostile. While many of them maybe tied to multiple engine parts.
Linear Transformations - Transformations That Are Usually Considered 1D and sometimes 2D such as Rotations about a single point or single line or axis.
Translations - Walking, running etc.
Scaling - Stretching or Skewing objects
Rotation - Rotating with an angle about a single axis, vertex(point) or vector(line).
Quadratic Transformations - Throwing a Ball, Jumping etc.
Affine Transformations - Matrix Transformations going from one View to another such as going from model to view to projection and there are others such as orthographic view; especially going from a 3D structure and to transform the data to fit onto a flat 2D plane such as your screen without the presence of stretching, tearing, and artifacts.
Complex Transformations - Rotations done in 3D among multiple axis of rotations at a time. Using Euler Angles this has a side effect called Gimbal Lock; to avoid it, we use Quaternions. Which is 4D mathematics involving matrices and complex numbers.
Scene Graph - Bringing them together to produce a working scene of related objects that defines the current world with all of its parts and where one object is in position related to another. There are many types of these graphs: BSP, Root - Node with leafs, Marching Cubes, Cubic or Octal Space Partitioning systems, advanced hashing table techniques etc.
The HUD - Heads Up Display
GUI-System - Graphical User Interface (Buttons, List Boxes, Text Boxes etc.)
Text Rendering System - How different fonts will be rendered.
Physics Engine - The Engine that keeps the game's refresh rate of its drawing calls to be in sync with both the monitors refresh rate and the GPU's drawing rate. It also does all the basic calculations for the majority of general motion that can be applied to any non static game object. It is also used in doing collision detections, ray tracing, and used for doing animations.
Animation Engine - Usually a collection of character and object animations that are predefined to some default value yet may be modified by a scripting language if such a scripting process is designed to be a part of the overall engine. (Better know how to write a parser and a compiler if you decide to include your own scripting language)
Audio Engine - Usually the controller that knows when to play the expected audio file either it be a background music, transitional music, or some sound effect. Also the ability to apply 3D Audio. Audio with multiple channels to apply specific sounds to a specific direction or location. Also the ability to not just play audio from a file, but from a Stream Object.
AI Engine - The engine that controls the abilities of all NPCs (Non Playable Characters). Their paths, their goals, their hostility levels, choices of attacks, the dialogue of conversations with the Player etc.
Networking Engine - The engine responsibility to creating network type data packets of needed game information and to transmit them over the network between a server and a client. The ability to both write and read(parse) the packets.
Shader Engine - Engine responsible for handling all of the Shaders
Memory Management System - Engine responsible for handling all of the games assets in memory.
Putting it All Together
One of the most sought out design of a versatile game engine in the modern industry is to have a working engine or set of engines that works independently from any Game Content, Game Properties, and Game Logic so that a Game Developer can rip out the old game with all of its source code and game assets and the Engine doesn't break to make another game using the same existing engine. Making it modular and portable and as generic as possible is another important aspect. Albeit there may be some modifications to the engine from one Game Style to another. Such as the difference between a First Person Shooter and an MMORPG. Special Care needs to be taken when beginning to work in this field for there are a lot of things to consider such as the topic of Batch Rendering.
Once you have a working 3D-2D Game Engine with all of its various parts that is nearly bug free and fully optimized then the next step is to get involved with your other applications for content creation where you will need to make your textures (or download free ware ones), make your models, create your levels or immediate map view of the world, create your sound effects and music or ( download free ware). Then plan out the game logic with a starting point, ending point and goals for different things in between. Of course you could also pay to buy some of that content or pay others to do some of the work for you. This is why it takes studios such as EA, Valve, Blizzard, Bethesda etc. a few years with thousands of people to make any decent 3D Game either it be for the Play Station, X-Box, or PC systems. But this is where to start! The Things to Consider and What it actually Takes to Make a Working 3D Game!
This is only the tip of the Iceberg too for there are many other concepts that I left out or didn't cover such as the different types of algorithms for doing specific tasks such as the order of rending objects base on distance from camera due to transparencies. I didn't even cover things such as error handlers, exception handlers, logging information to a file or screen during debugging and building modes before final release version, multi-threading, parallel programming, even working down to ASM instruction sets, concurrency, etc.
The bottom line is the most important thing you need into making a game is the dedication to learn and then to try and fail, fail, fail and fail again, but to get back up and try again until you accomplish it. For this you will need 1,000s and 1,000s of Hours! Researching, Programming, DEBUGGING, DEBUGGING, DEBUGGING! Then Repeat it all again! Why Repeat? Well once you get the hang of it; that technology is old, and you have to learn about the new technology and the new compilers, and the new APIs etc.!
If you want to know the inner workings of a basic complete 3D Game Engine from scratch check out this website MarekKnows and you can follow his series of video tutorials. His projects are primarily on windows using MSVS's IDE's and Compilers and he focuses on OpenGL. Once you have a grasp of the concepts of the Game Engine and later in a newer series Shader Engine, you can easily take those concepts and apply them to DirectX if you so choose, but you will have to change the Windows Code to support DirectX as opposed to OpenGL, and then learn how to use their API in place of OpenGL. The two are quite different!
StackOverflow etiquette means you'll get closed as "too broad" for a question. Specific questions are best here.
If you are mostly interested in "just writing a game", then you should consider using an existing engine like Unity 3D or Unreal Engine. Unity 3D is C#, which is a very common language in business development and is reasonably popular for mobile development. Unreal Engine is UnrealScript and C++ so is more common for console and PC development. Both have large, vibrant online communities. There are of course lots of other engines and game toolkits out there, so don't consider this an exhaustive listing.
If you are more interested in expanding your knowledge of computer programming through the effort of writing a game 'from scratch', or if you are hoping to become an expert in graphics and/or game technologies, then C++ is a better route. To get started with DirectX 11 in C++, you should look at DirectX Tool Kit which provides a good starting place for hacking together your own DirectX game engine.
If you find the C++ language barrier a bit too much, but want something more 'techy', then take a look at SharpDX .NET managed assemblies for using DirectX APIs. See DirectX and .NET.
For some book recommendations for learning DirectX 11, see this blog post. Be aware that many of the online tutorials (including the very popular Rastertek tutorials) are somewhat dated as they make heavy use of the legacy DirectX SDK. For details on that exactly that means when using Visual Studio 2015, see MSDN and Where is the DirectX SDK (2015 Edition)?.
C++ is a bit like the game of Go: It's relative easy to pick up the basics, but it takes years of experience to really master it. C++ provides an immense level of control, and most game engines are written in C++ at some level. You can integrate easily with the large collection existing C/C++-based libraries, and it's supported for almost all gaming platforms. All that said, the language is quite complex, and can be challenging to learn the best coding patterns as well as develop the robust debugging skills and discipline needed to be highly productive with it.
BTW, there's a project focused on making C++ easier to learn you might want to check out, the C++ Core Guidelines.
Within C++ is a smaller, simpler, safer language struggling to get out. -- Bjarne Stroustrup

Three.js: What's the upper limit for holding 60 FPS on an average desktop?

I'm currently working on a game using Three.js. I've been studying software engineering for four years and have been working professionally on backends for two, but I've barely touched on graphics aside from some simple Unity experimenting.
I currently have ~22,000 vertices and ~8,000 faces according to renderstats.js, and my desktop (above average) can't run it above 20 FPS. I'm using Lambert material as well as a single ambient light, so I feel like this isn't too much to ask.
With these figures in mind, is this the expected behavior for three.js rendering?
I would be pretty sure that is not end of the line and you are probably missing some possibilities for massive performance-improvements.
But just to give you some numbers first,
if you leave everything fancy away (including three.js) and just render an ultra-simple point-cloud with one fragment rendered per point, you can easily get to rendering 10-20 million (yes, million) points/vertices on an average GPU.
just with simple shapes and material, I already got three.js to render something in the range of 500k triangles (at 1080p-resolution) at 60FPS without problem. You can probably take those numbers times 10 for latest high-end GPUs.
However, these kinds of numbers are not really helpful.
Some hints:
if you want to debug your rendering-performance, you should first add some metrics. Renderstats is good, but I'd recommend integrating http://spite.github.io/rstats/ for this (see the example).
generally the choice of material shouldn't matter too much, the GPU is way more capable than most people think. It's more likely a problem somewhere else in the pipeline. EDIT from comment: In some cases, like hi-resolution displays with slow GPUs (think mobile-devices) this might be less true and complicated shader-code can slow down your site, but it might worth be looking at the other points first. As the rendering itself happens off-thread (so you can't measure it's duration using regular tools like the devtools-profiler), you can use the EXT_disjoint_timer_query-extension to get some information about what is going on on the GPU.
the number of drawcalls shouldn't be too high: three.js needs to do a single drawcall for every Mesh and Points-object rendered in the scene and too many objects are generally a far bigger problem than objects with lots of vertices. Reducing the number of drawcalls can be done by merging multiple geometries into one and making use of multi-materials, vertex-colors and things like that.
if you are doing postprocessing, the GPU needs to render every pixel on screen several times. This might as well massively limit your performance. This can be optimized by merging multiple postprocessing-passes into one (I admit, that'd be a lot of hard work..)
another problem could be on the JS side: you should use the profiler or timeline-view from the chrome devtools to see if maybe it's the javascript that is taking too much time per frame (shouldn't be more than 8-12ms per frame). I've been told there are ways to optimize the javascript-performance as well :)

Should I render all the frames every secound?

I'm programming a 2D game with HTML5 canvas. so it won't take a lot of time for rendering all the objects and frames.
I want to ask about both 2D and 3D games.
Suppose that I made a change or more on one of the objects.should I render all the objects when I only need to redraw that object? is rendering all objects and frames is the only option? and if the game was 3D, won't rendering take a lot of time? specially on slow internet connections
So [performance] depends only on the browser and computer of the user.
If you are making a game which is played simultaneously across several computers then the connection speed is important. Otherwise, the connection speed is mostly irrelevant (except as it relates to initially downloading your game resources). A better CPU/GPU, memory on the local computer will most affect your app performance.
About performance
In theory, you get better performance if you redraw only those game items which have changed since the last frame.
In practice, redrawing one game item often requires redrawing multiple other game items. You might want to redraw the player who moved, but in doing so you must also redraw the background plus any other game items colliding with the player.
As a result, it's often better to clear the entire canvas and redraw all the game items in their current position. So the unsatisfying answer to your question is that you should test your unique game app to determine if you can efficiently redraw just changed items or if a complete redrawing is faster.

kineticjs layers redraw optimization

background: i am developing a real-time multiplayer html5 canvas game using kineticjs which primarily will be played on MOBILE PHONE BROWSERS. There's a lot going on in the game such as socket communication with the server every second, redrawing and animations using kineticjs based on the server response and all this on top of a heavy graphic interface. The game functions well in desktop browsers however is SLUGGISH on mobilephones. So, I need to find all the ways in which the code can be optimized.
My questions,
lets say I need to redraw a particular part of the screen based on a server response that I just received from server, should I keep these need-to-be-redrawn elements in a separate layer, so that I need to redraw fewer elements. As in my case I need to redraw ever second, will this lead to performance improvement?
If the answer to the above is yes, then what is the optimal number of layers in which I should divide my layout. I ask this because I have a lot of different parts on the screen that need to be redrawn based on different server responses (although not all at the same time), if all these need to be put in separate layers, I need to know how far I can stretch the logic above, for example can I have 10 different layers without sacrificing the performance which is the aim of all this exercise anyway.
Eric Rowell (creator of KineticJS) has done some stress tests here: http://www.html5canvastutorials.com/labs/html5-canvas-kineticjs-drag-and-drop-stress-test-with-1000-shapes/
And he says this:
"Create 10 layers each containing 1000 shapes to create 10,000 shapes. This greatly improves performance because only 1,000 shapes will have to be drawn at a time when a circle is removed from a layer rather than all 10,000 shapes."
"Keep in mind that having too many layers can also slow down performance. I found that using 10 layers each made up of 1,000 shapes performs better than 20 layers with 500 shapes or 5 layers with 2,000 shapes."
So your takeaway is that
1.Yes, multiple canvas layers which isolate different groups of redrawables is the way to go.
And,
2.To optimize the tradeoff ( overhead of multiple canvas’s vs simplicity of 1 canvas ), you need to test with your own particular code in the environments they will be operating within.
Good luck with your game :)

Resources