RobotCaleb / sudokuSolver_Optimized

An optimised multi-threaded backtracking sudoku solver, that only iterates for the narrowed down possibilities, includes usual simple unoptimized backtracking too

An Optimized Sudoku Solver

An optimised backtracking sudoku solver, that only iterates for the narrowed down possibilities, includes usual unoptimized backtracking too

Implements multi-threading, using upto 9 threads during optimisation, for each row.

And, if you have any questions about this or related, even what may seem dumb to you, feel free to start an issue, or dm me on twitter 🚀

Building

---

Windows

Two Ways ->

1. If you have MSVC compiler, open vs code from Visual Studio Developer Command Prompt, and click Ctrl + Shift + B to run the default build task

2. Using CMAKE,

   cd build
   cmake ..
   cmake --build .

   * Or open the .sln solution file using visual studio then get executable

   
   

   Using either way, you will get a sudokuSolver.exe, on a succesful build, that's it ! 😄

Linux

• Using CMAKE and makefiles-

  cd build
  cmake ..
  make .

  This will provide you with an executable named sudokuSolver, on a successful build 👍

Fun Fact - This time i have more build methods for Windows, not Linux... quite rare 😅

Working -

• The sudoku class (sudoku.hpp) is the primary class, and uses box<> (box.hpp) as elements of 9x9 grid (for now, they just store a single value)

• The sudoku::board is the grid storing the actual values... and EMPTY is an enum var equating to 0. sudoku::_possibilities is also a 2D array, storing std::set of possible values at that index

• sudoku::optimise iterates through each row, assigns possibile values to each cell... then iterate through each column, and then each 3x3 block, thereby reducing the possibilities, and hence the recursion depth

• sudoku::solve applies simple backtrakking to solve the board, if sudoku::optimise has been called earlier, then the optimised vairant of sudoku::backtrackSolve_opt will be used, else sudoku::backtrackSolve

If you want more detail about something else, or more briefing on the above ones, do raise an issue :D

Further Improvements -

• sudoku::optimise() can be populated with more sudoku tricks, like naked triplets/pairs and others.

  You may refer https://bestofsudoku.com/sudoku-strategy for some strategies