Subversion Repositories slepc-dev

Rev

Go to most recent revision | Show changed files | Directory listing | RSS feed

Filtering Options

Rev Age Author Path Log message
2016 1042d 13h eromero /branches/ Deleted the branch davidsones
1986 1064d 14h eromero /branches/ - Updated davidsones from trunk up to revision 1985
1985 1064d 14h eromero /branches/ - Renamed EPS_DAVIDSONES to EPS_DAVIDSON
- Changed davidsones.* davidson.*
- Removed the custom fortran wrapper tags from the EPSGDGet*
1984 1064d 14h eromero /branches/ - Enforced fixed target
- When restarting, the eigenvalues are not recalculated
- Minor bug at "EPSGDGetKrylovStart"
1982 1067d 08h eromero /branches/ - Added the create for the virtual DAVIDSON solver
- Fixed some minors bugs
1981 1068d 05h eromero /branches/ - bug fixed: DVD_NORMALIZE_UV does not compile in complex
- removed the warnings in complex
- bug detected: JD does not work properly in complex
- bug fixed: minor bug in dvd_improvex_jd
1980 1068d 06h eromero /branches/ - Include EPSGD???InitialSize and -eps_gd_initial_size replacing -eps_davidsones_initv
- Removed the experimental macros EPSObj{Get,Set}{Prop,2Props} and replacing by functions in traditional PETSc manner
- Added 3 projectors more
- Removed dvd_improvex_{gdbasic,gdolsen}. The GD is implemented by dvd_improvex_jd with ksp==0. Advantages: less code and GD can use the JD projectors
1978 1071d 07h eromero /branches/ - bug fixed: PetscLogStageRegister is only allowed to be called once
1977 1071d 09h eromero /branches/ - davidson makefile file missing
1976 1071d 09h eromero /branches/ - Transformation of the experimental Davidson solver "DAVIDSONES" to a full integrated family of solvers "davidson". The first solver is "gd". The code shared by the "davidson" solvers is packaged by the virtual solver "DAVIDSON": EPSSetUp_DAVIDSON, EPSSolve_DAVIDSON...
- Partially fixed bug: the size of the starting searching subspace is set by eps->nini, but this value can be negative.
- Integration of "davidsones_harm" as "eps_extraction"
1975 1076d 09h eromero /branches/ - Updated from trunk up to r1974
- Made davidsones compatible with EPSSetInitialSpace
1972 1084d 10h eromero /branches/ - Now dvd_updateV_{update,conv,restart}_gen don't call to other of them
- Bug fixed: left subspace adapted to converge pairs of conjugate eigenvalues
1969 1090d 17h eromero /branches/ - Bug fixed: dvd_improvex_get_eigenvectors used the auxiliar memory d->auxS, instead of the local available memory
1968 1091d 08h eromero /branches/ - Added a routine for computing the associated eigenvector to a Schur decomposition
- Bug fixed: now nR starts with PETSC_MAX
1967 1096d 09h eromero /branches/ - fixed bug: minor error at EPSComputeVectors_QZ
1966 1096d 10h eromero /branches/ - In real arithmetic, the conjugate complex eigenpairs are enforced to converge together
1965 1096d 14h eromero /branches/ - bug fixed: residual vector in complex case but real arithmetic is computed with normalized eigenvectors
- bug fixed: a ksp doesn't allow operators of different size; so VecComp can operate with a virtual size
- bug fixed: RJDQZ needs that JD correction equation operates with eigenvectors instead of Schur vectors. For that the projectors of JD correction equation can operate with any projection of the left and right subspaces
- bug fixed: block size adapted to RJDQZ
- Added an UpdateVectors routine for non-contiguous stored vectors
- bug fixed: enforced restarting when improveX doesn't add any vector
- bug fixed: recomputing eigenvalues adapted to RJDQZ
- bugs fixed in SlepcUpdateVectorsS and SlepcUpdateStrideVectors
- lots of minor bug fixed...
1964 1104d 09h jroman /branches/ Fix output of ex13
1962 1104d 10h eromero /branches/ - Bug repaired: in EPSComputeVectors_QZ, tgevc sometimes fails computing the eigenpairs in the converged partial Schur decomposition
- Bug repaired: now dvd_updateV_conv_gen didn't converge more than ncv pairs
1960 1105d 14h eromero /branches/ - First version of RJDQZ

Show All