TSAdapt: Refactor interface and implementations* Move the handling of step limits, clipping factors, and safety factors to the base TSAdapt type to simplify implementation of subtypes.* Add TSAd
TSAdapt: Refactor interface and implementations* Move the handling of step limits, clipping factors, and safety factors to the base TSAdapt type to simplify implementation of subtypes.* Add TSAdapt{Set|Get}Clip() and remove TSAdaptBasic{Set|Get}Clip().* Add TSAdapt{Set|Get}Safety() to set safety factors.* Add TSAdaptGetStepLimits().
show more ...
Merge branch 'dalcinl/fix-bdf-adapt' into barry/update-petscfileretrieve
Allow timestep to be cut by adaptor after it is cut by failed solve;Commit-type: bug-fixFunded-by: IMEXProject: chemistryThanks-to: Emil Constantinescu <emconsta@mcs.anl.gov>
TSAdapt: Refactor handling of the always_accept flag* Add TSAdaptSetAlwaysAccept()* Store the always_accept flag in the base TSAdapt type* Fix/refactor bad printf args in TSAdaptView()* Fix TSAd
TSAdapt: Refactor handling of the always_accept flag* Add TSAdaptSetAlwaysAccept()* Store the always_accept flag in the base TSAdapt type* Fix/refactor bad printf args in TSAdaptView()* Fix TSAdaptSetFromOptions() to use TSAdaptSetXXX()
Very crude TSAdaptBasic change to prevent it from decreasing time-step immediately after it has increased itStill not completely satisfactory behavior for tchem behavior on gasoline model from Open
Very crude TSAdaptBasic change to prevent it from decreasing time-step immediately after it has increased itStill not completely satisfactory behavior for tchem behavior on gasoline model from OpenFoamTSAdapt monitor now always shows SNES failure reasonCommit-type: featureFunded-by: IMEXProject: ChemistryTime: 4 hours
Remove the use and definition of __FUNCT__ throughout the codeSince all modern C/C++ compilers provide this functionality we no longer need to provide it manually in PETScTime: 1.5 hoursThanks-t
Remove the use and definition of __FUNCT__ throughout the codeSince all modern C/C++ compilers provide this functionality we no longer need to provide it manually in PETScTime: 1.5 hoursThanks-to: Andreas Mang <andreas@ices.utexas.edu>
FIX TS adapt for the global error and to use atol and rtol separatelly
Extended the parameter list for the TS error norms to return the norms with respect to atol and rtol
Added new TS eror norms that use the error directly, added new example ODE in ex31, fixed interface comments/doc. Fixed TSAdapt glee.
FIX adapt error when atol and rtol are both zero; fixed order for glee wen used in adapt
FIX TSAdapt glee error computation and comments in ex31
FIX error in TSAdapt glee and renaming variables
modified TSGetTime error to include an index that indicates when to estimate the error.
extended the TS weighted error norm to compute the norms relative to absolute and relative tolerances separatelly
Setting up TSAdaptGlee
moving and renaming files
TS starting new GLEE adaptor
TSAdapt CFL: re-add guard for error message when !always_acceptThe CFL implementation was never completed due to a mathematicalgenerality issue. We still need to agree on a "good enough" solution
TSAdapt CFL: re-add guard for error message when !always_acceptThe CFL implementation was never completed due to a mathematicalgenerality issue. We still need to agree on a "good enough" solutionand finish the implementation, but until then, make sure it errors witha useful message so that people don't (a) think it's expected to work or(b) remove the error message without making it work.
TSAdapt: Improve error checking
TSAdapt: Add TSAdaptBasic{Set|Get}Clip()
TSAdapt: Rework implementation* Simplify handling of TSAdapt in TSALPHA* Remove broken adaptivity in TSTHETA and use the same backward difference aproach from TSALPHA* Update TSEULER to use TSA
TSAdapt: Rework implementation* Simplify handling of TSAdapt in TSALPHA* Remove broken adaptivity in TSTHETA and use the same backward difference aproach from TSALPHA* Update TSEULER to use TSAdapt (TODO: rejections and adaptivity)* Update TSSSP to use TSAdapt (TODO: rejections and adaptivity)
TSAdapt: Add and alternative API for weighted LTE evaluationLTE evaluation based on TSEvaluateStep() makes sense as long as themethod contains a lower/higher-order embedded scheme. However, for so
TSAdapt: Add and alternative API for weighted LTE evaluationLTE evaluation based on TSEvaluateStep() makes sense as long as themethod contains a lower/higher-order embedded scheme. However, for someother methods (e.g. BDF schemes), LTE evaluation is kind of more direct(based on history rather than difference of two solutions with differentorder). I'm adding an alternative API for TSAdapt to ask the TS subtypefor the weighted LTE. For now, this API is not exposed to users, howeverit would make sense to do so, as TSEvaluateWLTE() would be useful formonitoring error when running with fixed time step size.
Begin renaming the PetscOptions objects for the future where PetscOptions will not be a singleton
on solaris in TSAdaptChoose_Basic() negative root for zero could produce -infinity thuscausing time step to be decreased even though the error estimate was zero
fix: ex40,ex41 errors in src/ts/examples/tutorials/ on bsd with c++/complexPerhaps this code nees a separate section when 'enorm == 0) - as pow(0,-0.33) = inf.
123456