HYPRE is a software library of high performance preconditioners and solvers for the solution of large, sparse linear systems of equations. It features multigrid solvers for both structured and unstructured grid problems.

GEOS is a simulation framework for modeling coupled flow, transport, and geomechanics in the subsurface. The code provides advanced solvers for a number of target applications, including

HYPRE is a software library of high performance preconditioners and solvers for the solution of large, sparse linear systems of equations. It features multigrid solvers for both structured and unstructured grid problems.

SuiteSparse is a suite of sparse matrix algorithms, including: UMFPACK, multifrontal LU factorization; CHOLMOD, supernodal Cholesky; SPQR, multifrontal QR; KLU and BTF, sparse LU factorization, well-suited for circuit simulation; ordering methods (AMD, CAMD, COLAMD, and CCOLAMD); CSparse and CXSparse, a concise sparse Cholesky factorization package; and many other packages.

This package contains all of the above-mentioned parts.

This package provides the CUDA compiler and the CUDA run-time support libraries for NVIDIA GPUs, all of which are proprietary.

This package provides the CUDA compiler and the CUDA run-time support libraries for NVIDIA GPUs, all of which are proprietary.

This package provides the CUDA compiler and the CUDA run-time support libraries for NVIDIA GPUs, all of which are proprietary.

This package provides the CUDA compiler and the CUDA run-time support libraries for NVIDIA GPUs, all of which are proprietary.

This package provides the CUDA compiler and the CUDA run-time support libraries for NVIDIA GPUs, all of which are proprietary.

HDF5 is a suite that makes possible the management of extremely large and complex data collections.

HLIBpro is a C++ library implementing H-matrix arithmetic and other algorithms in the context of H-matrices, e.g. geometrical and algebraic clustering, visualisation, BEM discretisation. The focus of the implemention in HLIBpro is on robust algorithms on parallel computers. Here robustness is meant to apply to the algorithms itself as well as on the implementation.

A collection of libraries intended to be widely useful, and usable across a broad spectrum of applications.

The GNU Scientific Library is a library for numerical analysis in C and C++. It includes a wide range of mathematical routines, with over 1000 functions in total. Subject areas covered by the library include: differential equations, linear algebra, Fast Fourier Transforms and random numbers.

This package provides Intel's non-free compilers, icc and ifort. You may need to set the INTEL_LICENSE_FILE environment variable to use it.

Consider using GCC or Clang instead.

This package provides Intel's non-free compilers, icc and ifort. You may need to set the INTEL_LICENSE_FILE environment variable to use it.

Consider using GCC or Clang instead.

This package provides Intel's non-free compilers, icc and ifort. You may need to set the INTEL_LICENSE_FILE environment variable to use it.

Consider using GCC or Clang instead.

Medit was developped to visualize numerical simulation results on unstructured meshes in two and three dimensions. Scalar, vector and tensor fields can be easily associated and displayed with meshes.

ParMETIS is an MPI-based parallel library that implements a variety of algorithms for partitioning unstructured graphs, meshes, and for computing fill-reducing orderings of sparse matrices. ParMETIS extends the functionality provided by METIS and includes routines that are especially suited for parallel AMR computations and large scale numerical simulations. The algorithms implemented in ParMETIS are based on the parallel multilevel k-way graph-partitioning, adaptive repartitioning, and parallel multi-constrained partitioning schemes developed in our lab.

ParMETIS is an MPI-based parallel library that implements a variety of algorithms for partitioning unstructured graphs, meshes, and for computing fill-reducing orderings of sparse matrices. ParMETIS extends the functionality provided by METIS and includes routines that are especially suited for parallel AMR computations and large scale numerical simulations. The algorithms implemented in ParMETIS are based on the parallel multilevel k-way graph-partitioning, adaptive repartitioning, and parallel multi-constrained partitioning schemes developed in our lab.

ParMETIS is an MPI-based parallel library that implements a variety of algorithms for partitioning unstructured graphs, meshes, and for computing fill-reducing orderings of sparse matrices. ParMETIS extends the functionality provided by METIS and includes routines that are especially suited for parallel AMR computations and large scale numerical simulations. The algorithms implemented in ParMETIS are based on the parallel multilevel k-way graph-partitioning, adaptive repartitioning, and parallel multi-constrained partitioning schemes developed in our lab.

SuperLU_DIST is a parallel extension to the serial SuperLU library. It is targeted for distributed memory parallel machines. SuperLU_DIST is implemented in ANSI C, and MPI for communications.

ParMETIS is an MPI-based parallel library that implements a variety of algorithms for partitioning unstructured graphs, meshes, and for computing fill-reducing orderings of sparse matrices. ParMETIS extends the functionality provided by METIS and includes routines that are especially suited for parallel AMR computations and large scale numerical simulations. The algorithms implemented in ParMETIS are based on the parallel multilevel k-way graph-partitioning, adaptive repartitioning, and parallel multi-constrained partitioning schemes developed in our lab.

ParMETIS is an MPI-based parallel library that implements a variety of algorithms for partitioning unstructured graphs, meshes, and for computing fill-reducing orderings of sparse matrices. ParMETIS extends the functionality provided by METIS and includes routines that are especially suited for parallel AMR computations and large scale numerical simulations. The algorithms implemented in ParMETIS are based on the parallel multilevel k-way graph-partitioning, adaptive repartitioning, and parallel multi-constrained partitioning schemes developed in our lab.

This CUDA C++ Core Libraries (CCCL) unifies three essential CUDA C++ libraries into a single, convenient repository: Thrust, CUB and libcudacxx. The goal of CCCL is to provide CUDA C++ developers with building blocks that make it easier to write safe and efficient code. Bringing these libraries together streamlines your development process and broadens your ability to leverage the power of CUDA C++.