ViennaCL: Added smoothed aggregation preconditioner (saviennacl) similar to sacusp.Requires commit 6e09ddc from ViennaCL repository. Release 1.7.1 is insufficient.
Removed the use of PetscObjectQueryFunction() with with MatGetDiagonalBlock()that I missed when I first converted MatGetDiagonalBlock() from a query functionto a table functionTime: .3 hoursRe
Removed the use of PetscObjectQueryFunction() with with MatGetDiagonalBlock()that I missed when I first converted MatGetDiagonalBlock() from a query functionto a table functionTime: .3 hoursReported-by: nightly tests
show more ...
Merge branch 'maint'
Some Fortran stubs that returned strings could produce garbageSome Fortran stubs that returned strings did not properly use FIXRETURNCHAR() to blank out the end of the Fortran string. Thus the Fort
Some Fortran stubs that returned strings could produce garbageSome Fortran stubs that returned strings did not properly use FIXRETURNCHAR() to blank out the end of the Fortran string. Thus the Fortran string would contain "garbage".Time: .2 hoursReported-by: Mark Filipiak <m.filipiak@epcc.ed.ac.uk>
Merge branch 'stefano_zampini/fix-chebyshev-rhslocked'
PC: fix for complex builds. added PETSC_INTERN to PCPreSolveChangeRHS
Added PCPreSolveChangeRHS as suggested by Barry.the method is declared as intern in kspimpl.h to not introduce a new APIImplemented as a composable function to not change pcimpl.h
a variety of fixes for issues detected by the clang static analyzerReported-by: clang static analyzer
Removed PCSetInitialGuessNonzero() and PCGetInitialGuessNonzero();they make no sense since PCApply() is a operator and cannot depend on the initial guess.The routines were not used anyway. Also upd
Removed PCSetInitialGuessNonzero() and PCGetInitialGuessNonzero();they make no sense since PCApply() is a operator and cannot depend on the initial guess.The routines were not used anyway. Also updated the manual page for PCKSP developers notto clarify why there is not support for "passing the initial guess" into the inner KSPThanks-to: Dave May <dave.mayhem23@gmail.com>
removed the if (pc->setupcalled < 2) checks for calling PCSetUp() since the values are always less then 2Reported-by: Matthew Knepley <knepley@gmail.com>
cast
rename MatSetErrorIfFPE -> MatSetErrorIfFailure
add enum KSPConvergedSubReason
add PCFailedReason
Merge branch 'jed/process-options-handler' into barry/petscoptions-object
rename SemiRedundant Telescope
Merge branch 'master' into pr352/dmay/pctelescope/master
Merge branch 'master' into jed/process-options-handler
Options: pass PetscOptions* to options handlersPetscObjectAddOptionsHandler now takes an event handler that includes aPetscOptions argument. PetscObjectProcessOptionsHandlers requires thisas inp
Options: pass PetscOptions* to options handlersPetscObjectAddOptionsHandler now takes an event handler that includes aPetscOptions argument. PetscObjectProcessOptionsHandlers requires thisas input.
PC+SNES: Attach a level id to subsolvers in a multilevel solve- We should have done this years ago
PCSemiRedundant:* Added registration function and added dir to makefile
KSPCreateVecs() would under certain uncommon circumstances return incorrectly sized vectorsfor rectangular vectorsReported-by: Lars Kielhorn <lars.kielhorn@tailsit.com>PCApply() used the wrong m
KSPCreateVecs() would under certain uncommon circumstances return incorrectly sized vectorsfor rectangular vectorsReported-by: Lars Kielhorn <lars.kielhorn@tailsit.com>PCApply() used the wrong matrix to determine sizes for checking for vector length compatibility
doc fixes
Infrastructure that allows failures in PCSetUp(), PCApply(), MatMult() etc due to, for example a zero pivot or function evaluation outside its domain, to propagate up and become a KSP_DIVERGED inste
Infrastructure that allows failures in PCSetUp(), PCApply(), MatMult() etc due to, for example a zero pivot or function evaluation outside its domain, to propagate up and become a KSP_DIVERGED instead of generating an error that stops the program. In response to Issue 96. This includes failures in MatCreateSNESMF() applications due to domain errors. The mechanism to propagate some errors is by setting Info or Nan into the output vector and using the norm or inner product reductions in SNES or KSP to propagate the error condition to all processes and then handling them immediately after the norm or inner product.This allows, for example, ODE integrators to try again with a smaller time-step if the PCSetUp failed instead of requiring a complete restart of the run with other options.Currently some error conditions, such as function domain error in a line search may not get progated up using the correct SNESConvergedReasonSee src/snes/examples/tests/ex69.c for the handling of several conditions
PetscHeaderCreate: Shorten argument list and optimize memory allocation* Remove type arguments from the macro* Use a single memory allocation call instead of three
12345678910>>...23