We present the PSPIKE+ family of parallel hybrid SPARSE linear system solvers.
We propose the G-PAVER reordering for effective central banded preconditioners.
We develop a hybrid MPI/OpenMP based parallel implementation of PSPIKE+.
We demonstrate the robustness of PSPIKE+ on miscellaneous sparse linear systems.
We show the good parallel scalability of PSPIKE+ on distributed-memory architectures.