This project is aimed at implementing sudoku parallel solver in nVidia CUDA.
Clone this repo and change your cuda location in main folder. While in main folder, type:
$ mkdir bin/ && make
in order to have parallel solver compiled. If you want to compile sequential solver, type:
$ cd sequential/ && mkdir bin/ && make
Command line syntax:
$ ./parallel_sudoku_solver path_to_sudoku ...
There are example data in data/
folder. If you want to run sudoku solver on all example data, type:
$ cd Sudoku-parallel-solver/ && ./run.sh
Simillarly for sequential solver:
$ cd Sudoku-parallel-solver/sequential/ && ./run.sh
In case of out_of_memory
errors, you can control amount of memory used to run the program by changing kNBoards
variable inside include/device_resource_manager.cuh
. By default, the algorithm uses 2^22 sudoku boards.
TODO