Gmm++ User Documentation

• Introduction
• How to install and use GMM++
• Matrix and Vector type provided by GMM++
• • dense vectors
  • sparse vectors
  • skyline vectors
  • generic row and column matrices
  • dense matrices
  • sparse matrices
• Input and output with Harwell-Boeing and Matrix Market formats
• sub-vectors and sub-matrices
• • row and column of a matrix
• miscellaneous methods
• Basic linear algebra operations
• • scale and scaled
  • transposition
  • imaginary and real part
  • conjugate
  • add
  • mult
  • norms
  • trace
  • scalar product
• Solving triangular systems
• Dense LU decomposition
• Dense QR factorisation, eigenvalues and eigenvectors
• Condition number estimation
• Iterative solvers
• • iterations
  • various solvers
  • various preconditioners
  • Additive Schwarz method
• Catch errors
• Interface with BLAS, LAPACK or ATLAS
• Interface with SuperLU
• How to use GMM++ with QD type (double-double and quad-double)
• First steps with GMM++
• • How can I invert a matrix ?
  • How can I solve a linear system ?
  • How can I transform a vector into a matrix and reshape it ?
  • What is the better way to resize a matrix ?
• Deeper inside GMM++
• • The linalg_traits structure
  • How to iterate on the components of a vector
  • How to iterate on a matrix
  • How to make your algorithm working on all type of matrices
• How to disable verifications
• References
• Index

References

 [1]

L. F. Pavarino, Domain decomposition algorithms for the p-version finite element method for elliptic problems, Luca F. Pavarino, PhD thesis, Courant Institute of Mathematical Sciences, 1992.

---