EPSCISSStrategy

The strategy to be used when performing linear solves associated with integration points in the EPSCISS solver.

Synopsis

typedef enum { EPS_CISS_STRATEGY_USEST      = 1,
               EPS_CISS_STRATEGY_SPLIT      = 2,
               EPS_CISS_STRATEGY_MULTISHIFT = 3 } EPSCISSStrategy;

Values

• EPS_CISS_STRATEGY_USEST - use the ST object for the linear solves

• EPS_CISS_STRATEGY_SPLIT - use a collection of independent KSP objects

• EPS_CISS_STRATEGY_MULTISHIFT - use a KSP for multiple shifted systems such as KSPEKSM

Notes

The most computationally expensive step in EPSCISS is to solve \((A-z_i B)Y_i=BV\) for every integration point \(z_i\). This can be done in various ways.

Additional details are provided in the separate manual pages of each strategy.

The default is EPS_CISS_STRATEGY_USEST if the number of partitions is one (see EPSCISSSetSizes()) and EPS_CISS_STRATEGY_SPLIT otherwise.

See Also

EPS: Eigenvalue Problem Solver, EPSCISSSetStrategy(), EPSCISSSetSizes(), EPSCISSGetKSPs()

Level

advanced

Location

include/slepceps.h

---

Index of all EPS routines Table of Contents for all manual pages Index of all manual pages