DECOMPMETHOD

Description: Decomposition method:

PCGLSS	Selects the parallel sparse iterative solver available in all linear and nonlinear static solutions. This solver is recommended for large problems and will generally be faster than the VSS solver.
VSS	Selects the sparse direct solver available in all solutions. This solver is recommended for most problems. Significant performance degradation can occur if the RAM directive is set too low and an out of core solution is performed and/or physical memory is limited. The PCGLSS solver should be faster for these types of problems.
VIS	Selects the sparse iterative solver available in all except eigenvalue solutions. If VIS solver is selected for an eigenvalue solution, the VSS solver will be used. This solver is recommended for static solutions of models consisting mostly of solid elements. It can be significantly faster that the VSS solver in some cases and uses less resources (memory and disk space).
PSS	Selects the parallel sparse direct solver available in all solutions. This solver will be generally faster than the VSS solver especially on multiple CPU machines, but may require more memory.
AUTO	The program picks the best method based on the RAM directive setting, material properties, model size, and solution selected in the model. See also DECOMPAUTOSIZE.

Option/Type:

PCGLSS

VSS

VIS

PSS

AUTO

Default: AUTO

DECOMPMETHOD Directive Applicability Matrix

The matrix below depicts which initialization directives are applicable to the four linear equation solvers available. The DECOMPMETHOD directive is used to choose a particular solver.

Directive/Parameter	Solver (DECOMPMETHOD)
	PCGLSS (Sparse Iterative)	PSS (Sparse Direct)	PIS (Sparse Direct)	VSS (Sparse Direct)	VIS (Sparse Iterative)
DECOMPAUTOSIZE	✓	✓	✓	✓	✓
DECOMPMETHOD	✓	✓	✓	✓	✓
MAXSPARSEITER	✓				✓
RESEQGRIDMETHOD		✓	✓	✓
SPARSEITERTOL	✓		✓		✓
SPARSEITERMETHOD	✓
SPARSEITERMODE	✓
SPARSEMETHOD				✓