This program uses CUDA to compare CPU and GPU execution times .
cpu.cpp
g++ cpu.cpp -o cpu.out -std=c++11 -lstdc++
gpu.cu
nvcc gpu.cu -o gpu.out
cpu.out {square matrix size N} {Number of calculations}
./cpu.out 512 1
gpu.out {square matrix size N} {Number of calculations}
./gpu.out 512 1