Overview / Usage

CG is a popular method being used for solving systems of linear equations. This kind of problems can be found in different areas including simulations (FEM, SPICE) and machine learning. Purpose of this project is to implement CG and it's modified version named Preconditioned Conjugate Gradient method using Nvidia CUDA (maybe OCL in a future), compare CG and PCG (with different preconditioners) and perform overall benchmarking.