Eigen, cuSparse and SuiteSparse solvers

CMakeLists.txt

@@ -58,6 +58,8 @@ include(cmake/HandleEigen.cmake)            # Eigen3
 include(cmake/HandleMetis.cmake)            # metis
 include(cmake/HandleMKL.cmake)              # MKL
 include(cmake/HandleOpenMP.cmake)           # OpenMP
+include(cmake/HandleSuiteSparse.cmake)      # SuiteSparse sparse linear solver
+include(cmake/HandleCuSparse.cmake)         # cuSPARSE sparse linear solver
 include(cmake/HandlePerfTools.cmake)        # Google perftools
 include(cmake/HandlePython.cmake)           # Python options and commands
 include(cmake/HandleTBB.cmake)              # TBB

INSTALL.md

@@ -29,6 +29,8 @@ $ make install
        `GTSAM_WITH_EIGEN_MKL_OPENMP` to `ON`; however, best performance is usually
        achieved with MKL disabled. We therefore advise you to benchmark your problem 
        before using MKL.
+     - SuiteSparse is an alternate sparse matrix solver, but it is rarely faster than GTSAM's default variable elimination implementation.  Once enabled by toggling `GTSAM_WITH_SUITESPARSE`, it can be specified as the solver with `LinearSolverParams::linearSolverType`.
+     - cuSPARSE is the CUDA (GPU-accelerated) sparse matrix solver but is also rarely faster than GTSAM's default variable elimination implementation.  Once enabled by toggling `GTSAM_WITH_CUSPARSE`, it can be specified as the solver with `LinearSolverParams::linearSolverType`.
 
     Tested compilers:
 

README.md

@@ -50,6 +50,8 @@ Optional prerequisites - used automatically if findable by CMake:
 - [Intel Math Kernel Library (MKL)](http://software.intel.com/en-us/intel-mkl) (Ubuntu: [installing using APT](https://software.intel.com/en-us/articles/installing-intel-free-libs-and-python-apt-repo))
     - See [INSTALL.md](INSTALL.md) for more installation information
     - Note that MKL may not provide a speedup in all cases. Make sure to benchmark your problem with and without MKL.
+- [SuiteSparse](https://people.engr.tamu.edu/davis/suitesparse.html) (Ubuntu: `sudo apt-get install libsuitesparse-dev`)
+- [cuSPARSE](https://docs.nvidia.com/cuda/cusparse/index.html) (comes with [CUDA Toolkit](https://developer.nvidia.com/cuda-toolkit))
 
 ## GTSAM 4 Compatibility