An implementation of parallel Gauss-Jordan method in KJI form written in MPI C.
You could give your augmented matrix in a txt as an argument. If you don't then a random augmented matrix is generated. The "matrix_example.txt" in src folder is a txt file example.
We implement two variants of Gauss-Jordan method. The first seperates columns by zones and the second by card-shuffling.
In figures folder we give executions of both variations for p=2,4,8,16. As you can notice the executional times, the speedup and the efficiency for both variations follow the theoritical results.
We use MPI 3.1 and gcc 6.
In src folder run:
~$ mpicc suffle.c -o executable (or ~$ mpicc zones_col.c -o executable)
~$ mpirun -np number_of_processes ./executable aug_matrix.txt
These implementations were developed for the course of Parallel Algorithms in University.