AmgX provides a simple path to accelerated core solver technology on NVIDIA GPUs. AmgX provides up to 10x acceleration to the computationally intense linear solver portion of simulations, and is especially well suited for implicit unstructured methods.
It is a high performance, state-of-the-art library and includes a flexible solver composition system that allows a user to easily construct complex nested solvers and preconditioners.
Get Started with AmgX Today
The AmgX library offers optimized methods for massive parallelism, the flexibility to choose how the solvers are constructed, and is accessible through a simple C API that abstracts the parallelism and GPU implementation.
Using the methods and tools from the AmgX library, developers can easily create specialized solvers using AmgX core methods and rapidly deploy solution on GPU workstations, servers and clusters.
- Flexible configuration allows for nested solvers, smoothers, and preconditioners
- Ruge-Steuben algebraic multigrid
- Un-smoothed aggregation algebraic multigrid
- Krylov methods: PCG, GMRES, BiCGStab, and flexible variants
- Smoothers: Block-Jacobi, Gauss-Seidel, incomplete LU, Polynomial, dense LU
- Scalar or coupled block systems
- MPI support
- OpenMP support
- Flexible and simple high level C API
The AmgX Performance Benefit
Algebraic multigrid methods are applicable to a wide set of problems, and the AmgX library makes it easy to accelerate applications in various domains.
AmgX is now available as Open Source software at the AmgX repository.