MaWaTAM

mawatam is a simulator for the Maze-Walking aTAM, a model introduced in the paper Small tile sets that compute while solving mazes by M. Cook, T. Stérin and D. Woods. It is written in C++ and uses the SFML as graphic library.

Dependencies

We use the SFML, a powerful graphics, game and media library.

On Debian/Ubuntu, you can install SFML with: sudo apt install libsfml-dev, please refer to this guide for other systems.

Build

git clone
cd mawatam
mkdir build
cmake ..
make

Run

The simulator can either be run from a .yml input file (see input format below):

./mawatam -f examples/powers_of_two.yml

Or reading from stdin, in the same format:

python3 mawatam-tools/four_tiles_circuits.py prime_circuit 110  | ./mawatam -i

Note: The Maze-Walking aTAM model is defined in an asynchronous way. However, for convenience and because it did not change terminal assemblies in the case of the constructions we considered in the paper, the default growth mode of the simulator is synchronous. This can be changed by pressing G (see controls below).

Examples

You can simulate the constructions presented in the paper by running the following commands:

NAND-NXOR tile set

• python3 mawatam-tools/four_tiles_circuits.py prime_circuit 110 | ./mawatam -i
• python3 mawatam-tools/four_tiles_circuits.py crossover 01 | ./mawatam -i

Collatz tile set

• python3 mawatam-tools/Collatz_circuits.py prime_circuit_better 111 | ./mawatam -i
• python3 mawatam-tools/Collatz_circuits.py bridge_type_2_in_context 11 | ./mawatam -i
• python3 mawatam-tools/Collatz_circuits.py input_x_y_on_east_canonical_gate NAND | ./mawatam -i
• python3 mawatam-tools/Collatz_circuits.py input_x_y_on_east_canonical_gate NAND 01 | ./mawatam -i
• python3 mawatam-tools/Collatz_forward.py Collatz_forward 1001011 | ./mawatam -i
• python3 mawatam-tools/powers_of_two.py 100 | ./mawatam -i

Controls

• N: next simulation step
• R: reset simulation
• arrows: translate the scene
• mouse wheel pressed: translate the scene
• CTRL + mouse wheel down/up: zoom in/out
• Numpad + / Numpad -: zoom in/out
• = / -: zoom in/out
• G: change growth mode between asynchronous ordered, asynchronous random, synchronous
• A: prints information about the simulation and scene in the terminal
• D: dumps the current configuration in out.yml file

Input format

The input format is built on yaml.

# This input configuration constructs powers of two in base 3

# The format for specifying glues is:
# <glue_alphabet_name>.<glue_name>: <glue_strength>
# The alphabet name must comply with [_a-zA-Z][_a-zA-Z0-9]*
# The glue name must comply with [a-zA-Z0-9]
# Hence, the complete glue name must match [_a-zA-Z][_a-zA-Z0-9]*\\.[a-zA-Z0-9]
glues:
  bin.0: 1
  bin.1: 1
  ter.0: 1
  ter.1: 1
  ter.2: 1

temperature: 2

# The format for specifying tile types is:
# <tile_type_name>: [<glue_north_name>, <glue_east_name>, <glue_south_name>, <glue_west_name>]
# The tile type name must comply with [a-zA-Z0-9]
tileset_tile_types:
  0: [bin.0, ter.0, bin.0, ter.0]
  1: [bin.0, ter.1, bin.1, ter.0]
  2: [bin.0, ter.2, bin.0, ter.1]
  3: [bin.1, ter.0, bin.1, ter.1]
  4: [bin.1, ter.1, bin.0, ter.2]
  5: [bin.1, ter.2, bin.1, ter.2]

# When you specify a configuration, you can either use anonymous tile types that wont be registered in the tileset:
# <x>,<y>: [<glue_north_name>, <glue_east_name>, <glue_south_name>, <glue_west_name>]
# Or use tile types of the tileset:
# <x>, <y>: <tile_type_name>
configuration:
  0,0: [null, ter.0, null, null]
  0,1: [null, ter.0, null, null]
  0,2: [null, ter.0, null, null]
  0,3: [null, ter.0, null, null]
  0,4: [null, ter.0, null, null]
  0,5: [null, ter.0, null, null]
  0,6: [null, ter.0, null, null]
  0,7: [null, ter.0, null, null]
  0,8: [null, ter.0, null, null]
  0,9: [null, ter.0, null, null]
  0,10: [null, ter.0, null, null]
  0,11: [null, ter.0, null, null]
  0,12: [null, ter.0, null, null]
  0,13: [null, ter.0, null, null]
  0,14: [null, ter.0, null, null]
  0,15: [null, ter.0, null, null]
  0,16: [null, ter.0, null, null]
  1,-1: [bin.1, null, null, null]
  2,-1: [bin.0, null, null, null]
  3,-1: [bin.0, null, null, null]
  4,-1: [bin.0, null, null, null]
  5,-1: [bin.0, null, null, null]
  6,-1: [bin.0, null, null, null]
  7,-1: [bin.0, null, null, null]
  8,-1: [bin.0, null, null, null]
  9,-1: [bin.0, null, null, null]
  10,-1: [bin.0, null, null, null]
  11,-1: [bin.0, null, null, null]
  12,-1: [bin.0, null, null, null]
  13,-1: [bin.0, null, null, null]
  14,-1: [bin.0, null, null, null]
  15,-1: [bin.0, null, null, null]
  16,-1: [bin.0, null, null, null]
  17,-1: [bin.0, null, null, null]
  18,-1: [bin.0, null, null, null]
  19,-1: [bin.0, null, null, null]
  20,-1: [bin.0, null, null, null]
  21,-1: [bin.0, null, null, null]
  22,-1: [bin.0, null, null, null]
  23,-1: [bin.0, null, null, null]

# The two following entries are not implem yet

# You can specify the color with which the simulator will render
# glues that use a given alphabet name:
# <glue_alphabet_name>: #hexcolor or `auto` to let the simulator choose
# By default it is set to auto
glue_alphabets_color:
  bin: auto
  ter: auto

# The view will either display glue colors given their
# alphabet name or glue char can be toggled by pressing `e`
default_glue_color_mode: alphabet