Types

solver = CudssSolver(A::CuSparseMatrixCSR{T,INT}, structure::String, view::Char; index::Char='O')
solver = CudssSolver(rowPtr::CuVector{INT}, colVal::CuVector{INT}, nzVal::CuVector{T}, structure::String, view::Char; index::Char='O')
solver = CudssSolver(matrix::CudssMatrix{T,INT}, config::CudssConfig, data::CudssData)

The parameter type T is restricted to Float32, Float64, ComplexF32, or ComplexF64, while INT is restricted to Int32 or Int64.

CudssSolver contains all structures required to solve a linear system with cuDSS. It can also be used to solve a batch of linear systems sharing the same sparsity pattern. Two constructors of CudssSolver take as input the same parameters as CudssMatrix.

structure specifies the stucture for the sparse matrix:

• "G": General matrix – LDU factorization;
• "S": Real symmetric matrix – LDLᵀ factorization;
• "H": Complex Hermitian matrix – LDLᴴ factorization;
• "SPD": Symmetric positive-definite matrix – LLᵀ factorization;
• "HPD": Hermitian positive-definite matrix – LLᴴ factorization.

view specifies matrix view for the sparse matrix:

• 'L': Lower-triangular matrix and all values above the main diagonal are ignored;
• 'U': Upper-triangular matrix and all values below the main diagonal are ignored;
• 'F': Full matrix.

index specifies indexing base for the sparse matrix:

• 'Z': 0-based indexing;
• 'O': 1-based indexing.

CudssSolver can be also constructed from the three structures CudssMatrix, CudssConfig and CudssData if needed.

solver = CudssBatchedSolver(A::Vector{CuSparseMatrixCSR{T,INT}}, structure::String, view::Char; index::Char='O')
solver = CudssBatchedSolver(matrix::CudssBatchedMatrix{T,INT}, config::CudssConfig, data::CudssData)

The parameter type T is restricted to Float32, Float64, ComplexF32, or ComplexF64, while INT is restricted to Int32 or Int64.

CudssBatchedSolver contains all structures required to solve a batch of linear systems with cuDSS. One constructor of CudssBatchedSolver takes as input the same parameters as CudssBatchedMatrix.

structure specifies the stucture for the sparse matrices:

• "G": General matrix – LDU factorization;
• "S": Real symmetric matrix – LDLᵀ factorization;
• "H": Complex Hermitian matrix – LDLᴴ factorization;
• "SPD": Symmetric positive-definite matrix – LLᵀ factorization;
• "HPD": Hermitian positive-definite matrix – LLᴴ factorization.

view specifies matrix view for the sparse matrices:

• 'L': Lower-triangular matrix and all values above the main diagonal are ignored;
• 'U': Upper-triangular matrix and all values below the main diagonal are ignored;
• 'F': Full matrix.

index specifies indexing base for the sparse matrices:

• 'Z': 0-based indexing;
• 'O': 1-based indexing.

CudssBatchedSolver can be also constructed from the three structures CudssBatchedMatrix, CudssConfig and CudssData if needed.

matrix = CudssMatrix(::Type{T}, n::Integer; nbatch::Integer=1)
matrix = CudssMatrix(::Type{T}, m::Integer, n::Integer; nbatch::Integer=1)
matrix = CudssMatrix(b::CuVector{T})
matrix = CudssMatrix(B::CuMatrix{T})
matrix = CudssMatrix(A::CuSparseMatrixCSR{T,INT}, struture::String, view::Char; index::Char='O')
matrix = CudssMatrix(rowPtr::CuVector{INT}, colVal::CuVector{INT}, nzVal::CuVector{T}, struture::String, view::Char; index::Char='O')

The parameter type T is restricted to Float32, Float64, ComplexF32, or ComplexF64, while INT is restricted to Int32 or Int64.

CudssMatrix is a wrapper for CuVector, CuMatrix and CuSparseMatrixCSR. CudssMatrix can also represent a batch of CuSparseMatrixCSR sharing the same sparsity pattern. CudssMatrix is used to pass the matrix of the sparse linear system, as well as solution and right-hand side.

structure specifies the stucture for the sparse matrix:

• "G": General matrix – LDU factorization;
• "S": Real symmetric matrix – LDLᵀ factorization;
• "H": Complex Hermitian matrix – LDLᴴ factorization;
• "SPD": Symmetric positive-definite matrix – LLᵀ factorization;
• "HPD": Hermitian positive-definite matrix – LLᴴ factorization.

view specifies matrix view for the sparse matrix:

• 'L': Lower-triangular matrix and all values above the main diagonal are ignored;
• 'U': Upper-triangular matrix and all values below the main diagonal are ignored;
• 'F': Full matrix.

index specifies indexing base for the sparse matrix:

• 'Z': 0-based indexing;
• 'O': 1-based indexing.

matrix = CudssBatchedMatrix(b::Vector{CuVector{T}})
matrix = CudssBatchedMatrix(B::Vector{CuMatrix{T}})
matrix = CudssBatchedMatrix(A::Vector{CuSparseMatrixCSR{T,INT}}, struture::String, view::Char; index::Char='O')

The parameter type T is restricted to Float32, Float64, ComplexF32, or ComplexF64, while INT is restricted to Int32 or Int64.

CudssBatchedMatrix is a wrapper for Vector{CuVector}, Vector{CuMatrix} and Vector{CuSparseMatrixCSR}. CudssBatchedMatrix is used to pass the matrices of the sparse linear systems, as well as solutions and right-hand sides.

structure specifies the stucture for the sparse matrices:

• "G": General matrix – LDU factorization;
• "S": Real symmetric matrix – LDLᵀ factorization;
• "H": Complex Hermitian matrix – LDLᴴ factorization;
• "SPD": Symmetric positive-definite matrix – LLᵀ factorization;
• "HPD": Hermitian positive-definite matrix – LLᴴ factorization.

view specifies matrix view for the sparse matrices:

• 'L': Lower-triangular matrix and all values above the main diagonal are ignored;
• 'U': Upper-triangular matrix and all values below the main diagonal are ignored;
• 'F': Full matrix.

index specifies indexing base for the sparse matrices:

• 'Z': 0-based indexing;
• 'O': 1-based indexing.

config = CudssConfig()

CudssConfig stores configuration settings for the solver.

data = CudssData()
data = CudssData(cudss_handle::cudssHandle_t)

CudssData holds internal data (e.g., LU factors arrays).