pardiso iparm Parameter
This table describes all individual components of the
iparmparameter. Components which are not used must be initialized with 0. Default values are denoted with an asterisk (*).
Use default values.
are filled with default values.
You must supply all values in components
Fill-in reducing ordering for the input matrix.
You can control the parallel execution of the solver by explicitly setting the
MKL_NUM_THREADSenvironment variable. If fewer OpenMP threads are available than specified, the execution may slow down instead of speeding up. If
MKL_NUM_THREADSis not defined, then the solver uses all available processors.
The minimum degree algorithm [Li99] .
The nested dissection algorithm from the METIS package [Karypis98] .
The parallel (OpenMP) version of the nested dissection algorithm. It can decrease the time of computations on multi-core computers, especially when
PARDISO Phase 1 takes significant time.
= 3prevents the use of CNR mode (
> 0) because
PARDISO uses dynamic parallelism.
Reserved. Set to zero.
This parameter controls preconditioned CGS [Sonn89] for nonsymmetric or structurally symmetric matrices and Conjugate-Gradients for symmetric matrices.
has the form
The factorization is always computed as required by
CGS iteration replaces the computation of
LU. The preconditioner is
LUthat was computed at a previous step (the first step or last step with a failure) in a sequence of solutions needed for identical sparsity patterns.
CGS iteration for symmetric positive definite matrices replaces the computation of
T. The preconditioner is
Tthat was computed at a previous step (the first step or last step with a failure) in a sequence of solutions needed for identical sparsity patterns.
Lcontrols the stopping criterion of the Krylov Subspace iteration:
epsis used in the stopping criterion
|| / ||
|| < eps
|| = ||inv(
|| = ||inv(