Band SPD System
The Band SPD (Symmetric Positive Definite) system solver is optimized for symmetric positive definite matrices with a banded structure. It uses only half the band storage of Band General and specialized algorithms for SPD matrices.
Description
The Band SPD system: - Uses symmetric banded matrix storage - Optimized for symmetric positive definite matrices - Reduces memory usage by storing only half the band - Uses specialized Cholesky factorization
Parameters
Parameter |
Type |
Description |
---|---|---|
None |
This system solver doesn’t take any parameters |
Usage Example
# Create a Femora instance
# Create a Band SPD system solver
fm.analysis.system.create_system("BandSPD")
Notes
More efficient than Band General for SPD matrices
Uses approximately half the memory of Band General
Requires matrix to be symmetric positive definite
Performance depends on the quality of node numbering
Works well with RCM numbering scheme
Common choice for linear elastic analysis
Not suitable for non-symmetric or indefinite matrices