The P.U.R.G.E. Protocol

I recently participated at Ludum Dare 39, where you’re tasked with making a game, from scratch, in 48 hours.

The result was The P.UR.G.E. Protocol

In the game, you wake up in a room, resembling one of a spaceship. In there, a holographic computer is turned on, awaiting for your commands to start.. something. The current status for the ship is displayed next, showing that the ship may be badly damaged and hinting at a very complicated situation. And it is then when the computer starts asking you some questions.

But what is really going on?

Personally, it was an amazing experience. This was the first time I actually completed a game for a LD (I tried and failed a couple of times before) and it was a lot of fun. I’ll be doing a full post-mortem shortly.

You can find The P.U.R.G.E. Protocol either at the LDJam website or at The source code (yes, it’s open source) it’s available at Github.


The Ghost of Refactors Past…

…has come to hunt me once again. Although this time it’s not because of mistakes that I did. Instead, the problem lies in something that I missed completely.

Can you spot the problem in the following code?

void SomeObject::save( Stream &s )
   std::size_t count = getElementCount();
   s.write( count );

Well, it turns out std::size_t is NOT PLATFORM INDEPENDENT. And here’s the twist: I knew that since, well, forever, but I never paid any attention to it. That is, until it became a problem. And the problem was EVERYWHERE in the code.

First thing first. The C++ standard has this to say about std::size_t:

typedef /*implementation-defined*/ size_t;


What’s that supposed to mean? Basically, std::size_t may have different precision depending on the platform. For example, in a 32-bit architecture, std::size_t may be represented as a 32-bit unsigned integer. Something similar happens in a 64-bit platform.


std::size_t is supposed to help portability, right? That’s its whole purpose. And that’s true, of course. There’s nothing wrong with std::size_t itself.

Check the code above again. Go on, I’ll wait.

So, whenever we create a [binary] stream to serialize the scene, we can’t use std::size_t because it’s just plain wrong. It will be saved as a 32-bit integer in some platforms and 64-bit in others. Later, when the stream is loaded, the number of bytes read will depend on whatever the precision of the current platform is, regardless of what was used when saving. See the problem?

This means that we can’t share binary streams between different platforms because the data might be interpreted in different ways, leading to errors when generating scenes.

For the past few years, my main setup have been OS X and iOS, both 64-bit platforms. But one day I had to use streaming on 32-bit Android phones and, as you might have guessed by now, all hell break loose…

Entering crimild::types

I had to made a call here: either we can keep using std::size_t everywhere and handle the special case in the Stream class itself; or we can make use of fixed precision types for all values (specially integers) and therefore guaranteeing that the code will be platform independent.

I went for the second approach, which seems to me to be right choice. At the time of this writing, the new types are:

namespace crimild {

   using Int8 = int8_t;
   using Int16 = int16_t;
   using Int32 = int32_t;
   using Int64 = int64_t;

   using UInt8 = uint8_t;
   using UInt16 = uint16_t;
   using UInt32 = uint32_t;
   using UInt64 = uint64_t;

   using Real32 = float;
   using Real64 = double;

   using Bool = bool;
   using Size = UInt64;

As you can see, crimild::Size is always defined as a 64-bit unsigned integer regardless of the platform.

Yet that means I need to change every single type definition in the current code so it uses the new types. As you might have guessed, it’s a lot of work, so I’m going to do it on the fly. I think I already tackled the critical code (that is, streaming) by the time I’m writing this post, but there’s still much to be reviewed.

New code already makes use of the platform-independent types. For example, the new particle system is employing UInt32, Real64 and other new types and–

Oh, right, I haven’t talked about the new particle system yet… Well, spoiler alert: there’s a new particle system.

You want to know more? Come back next week, then 🙂




Progress Update – November 2016

This post summarizes all the things I’m currently working on, without any specific priority order (as usual).

Job System

Probably the biggest feature to be included in the next release (whenever that happens) is the new Job System.

Last year I implemented a Task Scheduler for asynchronous work and while it is (was?) useful, it does lack the mechanisms required for programs to properly work concurrently. In particular, there’s no way to wait for tasks to complete or group them together as one unit of work.

Enters the new Job System, in which we’re going to be able to schedule jobs (obviously), that may or may not be linked with other jobs (as in parent/child jobs) and wait for them to complete before continuing. A “work stealing” approach is being used internally and we can spawn as many worker threads as we need.

The new Job System will allow us to move to a more parallel architecture by implementing things like parallel visitors, multi-threaded render queues and more.

Surprisingly, the current state of this feature is quite advanced and looking good. I just pushed the main classes to the development branch and I expect to refactor the Simulation flow pretty soon.

Ray Tracer Improvements

I needed a way to test the new Job System before having to refactor the entire Simulation flow and the Ray Tracer seemed to be the perfect candidate.

Initial tests are promising, showing that the rendering time has been reduced by 60% (depending on the number of worker threads, of course). The image below took “only” 80 minutes to render in my Macbook using 8 worker threads.


I’ve included several fixes in the latest code and I’m going to move to phase two (actual geometry, lighting, textures, etc) probably before the end of the year.


As it is right now, the scripting system is mostly used for scene building, specifying objects and components in script files using Lua. But I want more. I need more. I want to be able to create components from Lua files and interact between them and first class ones written in C++. I want to be able to create a whole new simulation or game without enforcing the developer to write and compile C++ code.

The improved scripting tools are still in a design phase and even when I would like to include them in the next release, I know they might not make it.

Improving Audio Support

This is something that I had in my list for a while and I’m giving it a try as a side project. At the moment, Crimild supports audio clips using OpenAL (including positional audio, of course) only in WAV format and there’s no way to play background music.

My initial goal is to work with OGG files. By the end of phase one I should be able not only to play music files but also to support some simple mixing mechanisms.

I don’t expect this feature to be included in the next release, but it should be done early next year because it’s something that I need for a different, super secret, project.

Scene Streaming

Yes, this is back. You know, that super useful thing let you save and load entire scenes to disk, using a binary file format (which was there five years ago but then disappeared one day without reason).

A working implementation for scene streaming was included in the last version (I’m not kidding, it’s there). I’m constantly revisiting it by adding support for more components and entities. The new binary file format has proven quite good, and loading times have been reduced in a great amount.

That’s it. Well, not really. There might be one or two more things in my TODO list right now, but the ones above are the most important ones. I’m planning on having a new release before the year’s done, including most of these features. So, stay tuned.