Porting Graveyard Keeper to consoles and mobiles

right here blog publish, unless otherwise noted, turned into written through a member of Gamasutra's group.The strategies and opinions expressed are these of the creator and never Gamasutra or its mum or dad company.

in this article, I need to discuss our experience of porting our game "Graveyard Keeper" to mobiles and consoles.

Some basic counsel: Graveyard Keeper changed into developed with the aid of our team (Lazy undergo games) in solidarity after which ported to XBox One / PS4 / Nintendo change via our writer, tinyBuild. The cellular ports (iOS and Android) had been made in-apartment, by using us. also, a few months after the initial game unencumber, we made a DLC "Breaking dead" which was additionally ported to all systems. besides this, the game turned into continuously being patched and greater.

As you could see, the setup containing two dev groups isn't a straightforward one. mix it up with the task in the part of dynamic construction that has to be deeply optimized for some systems and also you'll get challenging technique.

Time to get technical.

taking a look at our online game you can doubtless suppose that it's an easy game to port, because it is second with pixel paintings. And if consoles are effective adequate to technique 4K/60fps, the pixel 2nd images wouldn't be an issue. Theoretically, that's true. however just about we used lots of very complicated solutions that have been not optimized for consoles and absolutely unoptimized for mobiles.

suppose of HDD velocity.

You probably heard that the current gen consoles structure is terribly comparable to the workstation. That's actual, however technically the hardware differs from the notebook. first of all, Xbox One and PS4 use complicated disk drives whereas most of modern PCs use SSD or hybrid drives. So, you have to keep in mind the loading instances.

think about our surprise, when the first construct turned into loading for all the way through essentially 5 minutes.

So, we had to rethink the total loading and asset storage methods. for instance, we began storing no longer the entire online game objects on the scene, but best the tips about them in fundamental (fake) objects. And after the loading is comprehensive we begun changing these fake objects with the true ones on the fly.

consider of multi-core.

also, in our event writing for the consoles requires the usage of all CPU cores at maximum making the code to work in parallel. via default, solidarity games are very based on a single core efficiency. And operating the online game "as is" might give you very delicate performance bottleneck when all different cores are idle whereas the simplest single core is below load.

We had to move distinct math and calculations to a unique thread operating in a parallel to free as a good deal core-#1 power as we are able to. truly, that's a superb element to bear in mind if you're no longer planning to port to consoles additionally because it will offer you a better performance anyway.

So, subsequent time we are able to plan every little thing we will to run in a separate thread/job. thankfully, solidarity had more suitable multi-threading a whole lot right through the last 12 months implementing job equipment and ECS.

consider of pictures optimization (even for 2d video games!).

a different heavy subject for us become the pictures. We had no previous journey optimizing anything for consoles or mobiles. for this reason we were ok to make a tree combined of 20 separate sprites just to make it less difficult to design. There became no need for such quantity of sprites for a single tree within the game itself.

Of course, we necessary to merge sprites to in the reduction of the volume of layers. that you can see the volume of preliminary geometry on the scene here:

The rightmost graphic suggests the "overdraw amount". That capability every time one sprite is drawn over one more, the color on that picture turns into brighter. So, that you could see that regardless of the simplicity of the seem to be, there are in reality loads of polygons drawn over every different.

additionally, all sprites of the game are lit through a posh shader that not handiest makes use of the regular-maps, but calculates the pseudo 3D-depth of the picture.

We had to make distinct models of visible optimization for distinct structures. as an example, we had to completely change the lights shader and its behavior for mobiles.

consider of platform-certain concerns.

also, there is loads of platform-particular work (specifically for consoles). for instance, you no longer simplest should guide the gamepad, however need to cope with distinctive wonderful instances like when a person plays the online game, the gamepad turns off, consumer connects a further controller and logs in to a distinct profile. And that's in the middle of the game. There are distinct giant lists of such situations that your online game should take care of.

a further concern you should think about is the saving equipment. Your video game needed to work with asynchronous saving routines and treat any file corruption may occur.

Patches and DLCs.

As outlined above, we labored with an external porting crew on the side of our publisher and pushing all new content, fixes and patches grew to become a difficult project. You should plan every thing forward. which you could't just tell the guys to push your remaining fix to Xbox if they are making a change port and each person is busy with that. 

And conserving in intellect that we're a small group that likes to enhance games with a excessive ardour, that doesn't go well with planning ahead. That's the purpose we couldn't without problems port the primary "Breaking dead" DLC to the Xbox. We didn't plan it very smartly. one day we simply aroused from sleep and determined to make zombies and put them within the first DLC. but guys at tinyBuild were busy deeply optimizing and porting the game to Nintendo swap. They couldn't drop that technique and begin to port this DLC correct away. They mandatory to conclude their port first.

That became the downside of the idea of constructing a simultaneous unlock on computer and Xbox. Ideally we should unlock a computer build, fix all bugs have been discovered and handiest then start porting it to consoles -- or spend months in isolated QA until now, with out real participant remarks. And simplest after the port is finished, beginning making a DLC. however we couldn't manage to pay for spending so a whole lot time on that. We had to do every little thing right away.

That's the motive we couldn't liberate the first Breaking dead DLC at no cost on other platforms (because it turned into on the workstation). when you improve on diverse structures, the prices and timelines ramp up relatively quickly. Even porting and releasing a patch eats up loads of resources.

Releasing titles, DLCs and updates all the time require passing platform compliance requirements, which means certification with the aid of platform holders. The greater complicated your video game mechanics and the greater the amount of the APIs you're the usage of the extra necessities you are going to should surpass. This potential that you simply need to examine your content against these requirements before submitting it for certification to make certain that the possibilities of passing it are good. in any other case you may additionally come across a big listing of considerations to fix and could no longer pass it (likely you just won't) inside first submission. extra submissions could also lead to further expenses. however the most harmful aspect about no longer passing the certification in time is delaying the introduced unencumber dates. 

here's extraordinarily problematic with quick iteration cycles we had on pc. it would be inconceivable to real-time merge new content material into the console builds, as proven in the diagram within the beginning of the article. Having in-apartment components will require at the least one programmer committed to each platform and a knowledgeable QA team that is continually testing the title being ported and record the concerns that fail the requirements so the porting developers can fix them before being found by the certification crew on the platform holder aspect. It requires building the whole porting structure inside the studio. And here's not most effective the components, but the certain potential and ability set worried in it.

lengthy story brief it's very effortless to unlock DLC on workstation and keep on iterating on it in true time, whereas launching the equal content material on consoles provides a lot of overhead expenses and complexity to the total method. 

Sviatoslav Cherkasovlead programmerat Lazy bear games