cuSOLVER

E760426

cuSOLVER is an NVIDIA GPU-accelerated linear algebra library that provides high-performance routines for solving dense and sparse systems of equations, eigenvalue problems, and related numerical tasks.

Jump to: Statements Referenced by

Statements (50)

Predicate Object
instanceOf GPU-accelerated linear algebra library
NVIDIA CUDA library
developer NVIDIA NERFINISHED
distributionForm shared library
static library
hasComponent cuSOLVER Dense (LAPACK-like) module
cuSOLVER RF (refactorization) module NERFINISHED
cuSOLVER Sparse module NERFINISHED
includedIn NVIDIA HPC SDK NERFINISHED
integratesWith Thrust NERFINISHED
cuBLAS NERFINISHED
cuFFT NERFINISHED
cuSPARSE NERFINISHED
languageBinding C NERFINISHED
C++
license proprietary NVIDIA license
optimizedFor GPU acceleration
high performance computing
partOf CUDA Toolkit NERFINISHED
programmingModel CUDA NERFINISHED
providesFactorization Cholesky factorization
LU factorization
QR factorization
SVD factorization
providesRoutineType dense linear algebra
eigenvalue problem solvers
least squares solvers
linear system solvers
matrix factorizations
sparse linear algebra
requires CUDA runtime NERFINISHED
CUDA-capable GPU
supportsDataType complex numbers
real numbers
supportsHardware NVIDIA GPU NERFINISHED
supportsMatrixFormat dense matrix
sparse matrix
supportsPrecision double precision floating point
single precision floating point
supportsSparseFormat CSC format GENERATED
CSR format GENERATED
block sparse formats GENERATED
targetDomain data analytics
engineering simulation
machine learning
scientific computing
useCase computing eigenvalues and eigenvectors
computing singular value decompositions
least squares fitting
solving linear systems Ax=b

Referenced by (2)

Full triples — surface form annotated when it differs from this entity's canonical label.

cuSPARSE integratesWith cuSOLVER
cuBLAS relatedLibrary cuSOLVER