Parallel Sparse Computation Toolkit

This paper presents a new software framework for solving large and sparse linear systems on current hybrid architectures, from small servers to high-end supercomputers, embedding multi-core CPUs Nvidia GPUs at the node level. The has modular structure is composed of three main components, which separate basic functionalities managing distributed matrices executing some matrix computations involved in iterative Krylov projection methods, eventually exploiting multi-threading CUDA-based programming models, setup application different types one-level multi-level algebraic preconditioners.