Jump to content

Multithreaded Architecture in Leadwerks Game Engine 5

Josh

2,531 views

Leadwerks Game Engine 5 is a restructuring of Leadwerks Game Engine 4 to adapt to the demands of virtual reality and leverage the full capabilities of modern and future hardware.

Basically, the main idea with VR is that if you don't maintain a steady 90 FPS in virtual reality, you will throw up.  Nausea may be the worst physiological feeling you can experience.  In fact, nausea has been rated by cancer patients as worse than pain.  Being sensitive to motion sickness myself, this is a problem I am very motivated to solve.

In a conventional renderer, running both your game logic and rendering at 60 hz (frames per second) seems reasonable.  However, when we up the framerate to the 90 hz required for fluid virtual reality experiences, it seems like an excessive demand on the game code.  Game logic normally handles AI, player input, and various other tasks, and those things don't have to be updated quite that often.

Distributing Tasks
The solution I have come up with is to decouple the game loop from the renderer.  In the diagram below, the game loop is running at only 30 hz, while the physics, culling, and rendering loops are running at independent frequencies.

multithread1.png.9a5b81c1fd8ac31eac590a913b23894c.png

Think of this as like gears on a bicycle.  Your pedals move slowly, but your wheels spin very fast.  The game logic loop is like your pedals, while the rendering loop is like the rear wheel it is connected to.

gears.jpg.2a420b3c735ed09ebaad61597144f6b8.jpg

Previously, your game logic needed to execute in about 8 milliseconds or it would start slowing down the framerate.  With my design here, your game code gets more than 32 milliseconds to execute, a lifetime in code execution time, while a steady framerate of 90 or 60 FPS is constantly maintained.

I actually came up with this idea on my own, but upon further research I found out this is exactly what Intel recommends.  The technique is called Free Step Mode.  The diagram below does not correspond to our engine design, but it does illustrate the idea that separate systems are operating at different speeds:

7951-2.jpg.a224d98361a9c2890ece33362afa653f.jpg

If all threads are set to execute at the same frequency, it is called Lock Step Mode.

7952.jpg.913116238ba12998a0b351549d732e16.jpg

Data Exchange
Data in the game loop is exchanged with the physics and navmesh threads, but is passed one-way on to the culling loop, where it is then passed in a single direction to the rendering loop.  This means there will be a slight delay between when an event occurs and when it makes its way to the rendering thread and the final screen output, but we are talking times on the level of perhaps 10 milliseconds, so it won't be noticeable.  The user will just see smooth motion.

multithread2.png.5f514bfdc821b4e6404455b91c10248a.png

Rather than risk a lot of mutex locks, data is going to be passed one-way and each thread will have a copy of the object.  The game loop will have the full entity class, but the rendering threads will only have a stripped-down class, something like this:

class RenderObject
{
public:
	Mat4 matrix;
	AABB aabb;
	std::vector<Surface*> surfaces;
};

The original entity this object corresponds to can be modified or deleted, without fear of affecting downstream threads.  Again, Intel confirmed what I already thought would be the best approach:

Quote

In order for a game engine to truly run parallel, with as little synchronization overhead as possible, it will need to have each system operate within its own execution state with as little interaction as possible to anything else that is going on in the engine. Data still needs to be shared however, but now instead of each system accessing a common data location to say, get position or orientation data, each system has its own copy. This removes the data dependency that exists between different parts of the engine. Notices of any changes made by a system to shared data are sent to a state manager which then queues up all the changes, called messaging. Once the different systems are done executing, they are notified of the state changes and update their internal data structures, which is also part of messaging. Using this mechanism greatly reduces synchronization overhead, allowing systems to act more independently.

-Designing the Framework of a Parallel Game Engine, Jeff Andrews, Intel
https://software.intel.com/en-us/articles/designing-the-framework-of-a-parallel-game-engine

But wait, isn't latency a huge problem in VR, and I just described a system that adds latency to the renderer?  Yes and no.  The rendering thread will constantly update the headset and controller orientations, every single frame, at 90 hz.  The rest of the world will be 1-2 frames behind, but it won't matter because it's not connected to your body.  You'll get smooth head motion with zero delays while at the same time relieving the renderer of all CPU-side bottlenecks.

Even for non-VR games, I believe this design will produce a massive performance boost unlike anything you've ever seen.



6 Comments


Recommended Comments

I don't understand, why render another frame if gameplay logic wasn't updated yet and nothing in the world changed? In non-VR games you will see two identical frames?

Will we be able to increase gameplay refresh rate? I think gameplay should be updated faster than rendering. My rhythm game feels much more responsive on 200 FPS than on 60 even though I have 60 hz monitor.

Share this comment


Link to comment
Just now, Genebris said:

I don't understand, why render another frame if gameplay logic wasn't updated yet and nothing in the world changed? In non-VR games you will see two identical frames?

Will we be able to increase gameplay refresh rate? I think gameplay should be updated faster than rendering. My rhythm game feels much more responsive on 200 FPS than on 60 even though I have 60 hz monitor.

The renderer can interpolate between two frames of data, creating a new frame in-between.

The frequencies of each system can probably be made adjustable.

Share this comment


Link to comment

Genebris,

While nothing in the game world may have changed, the camera may turn suddenly and a new rendered frame is necessary, one with out lag. I don't want game logic impacting movement of camera. 

 

 

Share this comment


Link to comment
1 minute ago, wayneg said:

Genebris,

While nothing in the game world may have changed, the camera may turn suddenly and a new rendered frame is necessary, one with out lag. I don't want game logic impacting movement of camera. 

 

 

That's the one point in the renderer where a mutex will lock, and a callback will be used to update the camera rotation and position.  The only thing that has to be instantaneous are the camera and VR controllers (if present).

Share this comment


Link to comment

I got goosebumps reading this. I am so excited to be able to leverage all of these great features. I cannot tell you how excited I am to have this the software in this architecture. It is really going to answer a lot of difficult questions for us in the engineering field as well. Well done leadwerks.

Share this comment


Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Add a comment...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Blog Entries

    • By Josh in Josh's Dev Blog 6
      You might have seen this graphic comparing the size of the world in different games. I've played Fuel, and never reached the end of the world in that game. You can drive for a very long time on those roads.

      We want to use the new engine for realistic simulations of air and ground movements. At normal cruising altitude of a commercial airliner, the pilot has a view range of about 400 kilometers. The image below shows that area (800 x 800 km). You can see the areas of the biggest games ever fit neatly into the corner of just our visible area.

      The gray space above is not the total world size, it is just the area you can see at once from high altitude. The total world size is about 50 times bigger.
      This is what I am working on now.
    • By Josh in Josh's Dev Blog 26
      Gamers have always been fascinated with the idea of endless areas to roam.  It seems we are always artificially constrained within a small area to play in, and the possibility of an entire world outside those bounds is tantalizing.  The game FUEL captured this idea by presenting the player with an enormous world that took hours to drive across:
      In the past, I always implemented terrain with one big heightmap texture, which had a fixed size like 1024x1024, 2048x2048, etc.  However, our vegetation system, featured in the book Game Engine Gems 3, required a different approach.  There was far too many instances of grass, trees, and rocks to store them all in memory, and I wanted to do something really radical.  The solution was to create an algorithm that could instantly calculate all the vegetation instances in a given area.  The algorithm would always produce the same result, but the actual data would never be saved, it was just retrieved in the area where you needed it, when you needed it.  So with a few modifications, our vegetation system is already set up to generate infinite instances far into the distance.

      However, terrain is problematic.  Just because an area is too far away to see doesn't mean it should stop existing.  If we don't store the terrain in memory then how do we prevent far away objects from falling into the ground?  I don't like the idea of disabling far away physics because it makes things very complex for the end user.  There are definitely some tricks we can add like not updating far away AI agents, but I want everything to just work by default, to the best of my ability.
      It was during the development of the vegetation system that I realized the MISSING PIECE to this puzzle.  The secret is in the way collision works with vegetation.  When any object moves all the collidable vegetation instances around it are retrieved and collision is performed on this fetched data.  We can do the exact same thing with terrain   Imagine a log rolling across the terrain.  We could use an algorithm to generate all the triangles it potentially could collide with, like in the image below.

      You can probably imagine how it would be easy to lay out an infinite grid of flat squares around the player, wherever he is standing in the world.

      What if we only save heightmap data for the squares the user modifies in the editor?  They can't possibly modify the entire universe, so let's just save their changes and make the default terrain flat.  It won't be very interesting, but it will work, right?
      What if instead of being flat by default, there was a function we had that would procedurally calculate the terrain height at any point?  The input would be the XZ position in the world and the output would be a heightmap value.

      If we used this, then we would have an entire procedurally generated terrain combined with parts that the developer modifies by hand with the terrain tools.  Only the hand-modified parts would have to be saved to a series of files that could be named "mapname_x_x.patch", i.e. "magickingdom_54_72.patch".  These patches could be loaded from disk as needed, and deleted from memory when no longer in use.
      The real magic would be in developing an algorithm that could quickly generate a height value given an XZ position.  A random seed could be introduced to allow us to create an endless variety of procedural landscapes to explore.  Perhaps a large brush could even be used to assign characteristics to an entire region like "mountainy", "plains", etc.
      The possibilities of what we can do in Leadwerks Engine 5 are intriguing.  Granted I don't have all the answers right now, but implementing a system like this would be a major step forward that unlocks an enormous world to explore.  What do you think?

    • By Haydenmango in Snowboarding Development Blog 6
      So I've been researching snowboarding lately to get an idea of what animations and mechanics I need to create for my game.  I have learned lots of interesting things since I've only seen snow once or twice in my entire life and have never even tried snowboarding or any other board sports (skateboarding, surfing, etc.) for that matter.
       
      Snowboarding tricks are quite interesting as they are mostly derived from skateboarding.  Snowboarding tricks pay homage to their equivalent skating tricks by sharing many concepts and names.  For example basic grabs in snowboarding share the same concepts and names as skateboarding: indy, mute, method, stalefish, nosegrab, and tailgrab.  Something interesting to note is in snowboarding you can grab Tindy or Tailfish but this is considered poor form since these grabs can't be done on a skateboard (due to the board not being attached to the skaters feet) and grabbing these areas is generally something a novice snowboarder does when failing or "half-assing" a normal grab.  Check out this diagram to see how grabs work -
       
       
      So, after reading lots of text descriptions for tricks I was still confused by what all these terms meant and how they were actually applied.  So my next step was to look up these tricks actually being done and I found some really cool videos showing off how to do various tricks.  This video in particular is the best reference material I've found as it contains nearly every trick back to back with labeled names and some tweaks -
       
      Sadly my rigged model doesn't handle leg animations with the snowboard that well so I can't animate as many tricks as I want to.  Regardless there will still be around 15 total grab/air tricks in the game.  Now it's time for me to stop procrastinating and start animating!  
×
×
  • Create New...