Multi-GPU aggregation-based AMG preconditioner for iterative linear solvers (2303.02352v1)

Abstract: We present and release in open source format a sparse linear solver which efficiently exploits heterogeneous parallel computers. The solver can be easily integrated into scientific applications that need to solve large and sparse linear systems on modern parallel computers made of hybrid nodes hosting NVIDIA Graphics Processing Unit (GPU) accelerators. The work extends our previous efforts in the exploitation of a single GPU accelerator and proposes an implementation, based on the hybrid MPI-CUDA software environment, of a Krylov-type linear solver relying on an efficient Algebraic MultiGrid (AMG) preconditioner already available in the BootCMatchG library. Our design for the hybrid implementation has been driven by the best practices for minimizing data communication overhead when multiple GPUs are employed, yet preserving the efficiency of the single GPU kernels. Strong and weak scalability results on well-known benchmark test cases of the new version of the library are discussed. Comparisons with the Nvidia AmgX solution show an improvement of up to 2.0x in the solve phase.