This is a small benchmark for linear solvers from the Eigen and amgcl library.
This version requires the linear solver MR from Eigen
Download and then run
cd build
cmake -DCMAKE_PREFIX_PATH="PATH_TO_AMGCL;PATH_TO_EIGEN" ..
make
If you want to use CUDA please add
-DUSE_CUDA=True
In build/src you will find a benchmark executable with -h you get help options.
For running it requires a sparse matrix A and a vector b in the MatrixMarket format.