GPU Project - Parallelization of Ford Fulkerson Algorithm

Serial and parallel versions of the code have been attached along with the files used to test correctness.

To generate testcases

g++ generate_test_cases.cpp -std=c++11 ./a.out > <output_file>

Modify the generate_test_cases.cpp to generate testcases with different number of vertices and edges.

To execute the serial implementation

g++ serial_code.cpp -std=c++11 ./a.out < <input_file>

To execute the parallel implementations

nvcc parallel_code_v1.cu
nvcc parallel_code_v2.cu
nvcc -arch=sm_35 -rdc=true parallel_code_v3.cu
./a.out < <input_file>