Nikolai Sakharnykh

Learn More
Modern scientific discovery is driven by an insatiable demand for computing performance. The HPC community is targeting development of supercomputers able to sustain 1 ExaFlops by the year 2020 and power consumption is the primary obstacle to achieving this goal. A combination of architectural improvements, circuit design, and manufacturing technologies(More)
In this presentation we discuss the AmgX iterative solve framework, which includes implementation of two forms of Algebraic Multigrid (AMG), Krylov subspace methods, as well as parallel variants of typical smoothers and preconditioners. The goal of AmgX is to fully utilize the hardware capabilities of large GPU-­‐ accelerated clusters, especially for(More)
  • 1