In this report, we present a novel three-dimensional multi-group radiation transport iterative solver—pARIS. pARIS is the abbreviation of "Parallel Advanced Radiation-transport Iterative Solver", whose aim is to test and develop advanced parallel iterative algorithms for nonlinear transport equations on complex three-dimensional unstructured grids. For traditional implicit discrete ordinate method (DOM), one generally eliminates the electron temperature variable first and will then obtain a pseudo-scattering term that couples the radiation intensity in different directions. A popular solution procedure for the discrete system is sweeping method embeded into source iteration. However source iteration may converge slowly and sweeping can not be successfully implemented on general unstructured grids. To mitigate these issues pARIS makes use of preconditioned Krylov subspace methods to solve the single-group equation instead of classical sweeping strategy. And to improve the convergence of source iteration, Gauss-Seidel’s iteration and Larsen's GTA method are both implemented. A large number of numerical experiments are conducted to show the parallel scalability, flexibility and robustness of pARIS.