Clients#
rocALUTION clients host a variety of different examples as well as a unit test package. For detailed instructions on how to build rocALUTION with clients, see Installation on Linux or Installation on Windows.
Examples#
The examples collection offers different possible set-ups of solvers and preconditioners. The following tables provide a quick overview of various examples:
Example |
Description |
---|---|
|
Algebraic Multigrid solver (smoothed aggregation scheme, GS smoothing) |
|
GMRES solver with Additive Schwarz preconditioning |
|
Asynchronous rocALUTION object transfer |
|
Benchmarking important sparse functions |
|
BiCGStab solver with multicolored Gauss-Seidel preconditioning |
|
GMRES solver with blockwise multicolored ILU preconditioning |
|
CG solver with Algebraic Multigrid (smoothed aggregation scheme) preconditioning |
|
CG solver with Jacobi preconditioning |
|
CG solver with ILU preconditioning using Cuthill McKee ordering |
|
Matrix inversion |
|
Flexible GMRES solver with multicolored Gauss-Seidel preconditioning |
|
Fixed-Point iteration scheme using Jacobi relaxation |
|
GMRES solver with multicolored Gauss-Seidel preconditioning |
|
Induced Dimension Reduction solver with Jacobi preconditioning |
|
Sparse matrix unique key computation |
|
CG solver with multi-elimination preconditioning |
|
Mixed-precision CG solver with multicolored ILU preconditioning |
|
CG solver using Chebyshev preconditioning and power method for eigenvalue approximation |
|
Sparse Matrix Vector multiplication |
|
BiCGStab solver with multicolored ILU preconditioning for saddle point problems |
|
CG solver using stencil as operator |
|
CG solver with Truncated Neumann Series preconditioning |
|
FGMRES solver with variable preconditioning |
Example (MPI) |
Description |
---|---|
|
Benchmarking important sparse functions |
|
BiCGStab solver with multicolored Gauss-Seidel preconditioning |
|
CG solver with Algebraic Multigrid (pairwise aggregation scheme) preconditioning |
|
CG solver with Jacobi preconditioning |
|
Flexible CG solver with ILU preconditioning |
|
Flexible GMRES solver with SParse Approximate Inverse preconditioning |
|
File I/O with CG solver and Factorized Sparse Approximate Inverse preconditioning |
|
IDR solver with Factorized Sparse Approximate Inverse preconditioning |
|
QMRCGStab solver with ILU-T preconditioning |
Unit Tests#
There are multiple unit tests available to test for bad arguments, invalid parameters, and solver and preconditioner functionality. These unit tests are based on google test. The tests cover a variety of solver, preconditioning, and matrix format combinations and can be performed on all available backends.