Implementation of the famous Conway's Game of Life simulation in C#, using the Unity game engine.
The project is currently deployed at https://conwayunity.azurewebsites.net/
Each cell is represented as a byte, with the top 3 bits unused, and next 4 bits used to store a count of the cell's live neighbours, and the lowest bit being used to store the cell's state (alive = high or dead = low).
i.e - XXXN NNNS
Where X is an unused bit; N is the cell's live neighbour count; S is the cell's state.
This implementation is an adapted version of Michael Abrash's optimised version of Life - see Listing 17.5 L17-5.CPP.
Profiling of the initial (very simple) implementation suggested 94% of CPU usage was down to the LiveNeighbourCount function (function used to count how many live neighbours a cell has).
This initial implementation resulted in ~20 - 25 generations / second.
The optimised implementation is heavily based on Michael Abrash's optimised version of the simulation. While not making use of pointer arithmetic and still using a 2D array with a wrapping index function, CPU usage by the main update loop has dropped dramatically (~11%), with the result now ~60 generations / second (possible limited by Unity's framerate?).
- The initial game state is random, and currently cannot be changed in any way. The probability of any given cell being alive at the start of the game is 0.08.
- No stoppage mechanism of any kind is implemented. The simulation will keep going even if there are no more cells alive or a stable equilibrium has been reached.
- Implementation details are found in the Scripts directory.