aboutsummaryrefslogtreecommitdiff
path: root/Carpet/CarpetReduce
Commit message (Collapse)AuthorAge
* CarpetReduce: Correct boundary sizes in reduction mask setupErik Schnetter2011-12-14
| | | | The previous commit had an error.
* Erik's patch to fix a reduction operator problem.Christian Reisswig2011-12-14
| | | | Without this fix, reductions may depend on the number of processors used.
* CarpetReduce: Rewrite prolongation and reduction mask handlingErik Schnetter2011-12-14
| | | | | Rewrite prolongation and reduction mask handling to use the pre-calculated information in the dh structure.
* CarpetReduce: Correct check for sufficient ghost zonesErik Schnetter2011-12-14
|
* CarpetReduce: #include <omp.h>Erik Schnetter2011-12-14
|
* CarpetReduce: Rewrite OpenMP parallelisation for PGI compilersErik Schnetter2011-12-14
| | | | | Special case the OpenMP parallelisation of the main reduction loop for PGI compilers to circumvent a compiler bug.
* CarpetReduce: declare aliased GetBoundarySpecification functionsRoland Haas2011-12-14
| | | | | from CoordBase and MultiPact in interface.ccl and by including cctk_Functions.h
* CarpetReduce: Correct weight calculation for cell-centred grid hierarchiesErik Schnetter2011-12-14
|
* CarpetReduce: Remove dependency on LoopControl; use NoMPIErik Schnetter2011-12-14
|
* CarpetReduce: Correct handling staggered boundariesErik Schnetter2011-12-14
|
* CarpetReduce: ReformatErik Schnetter2011-12-14
|
* CarpetReduce: Remove unused header fileErik Schnetter2011-12-14
|
* CarpetReduce: Adapt handling mesh refinement to changes in CarpetLibErik Schnetter2011-12-14
| | | | Completely rewrite the code, using the active and fine_active data provided by CarpetLib.
* CarpetReduce: Support multi-block systemsErik Schnetter2011-12-14
|
* CarpetReduce: Add more self testsErik Schnetter2011-12-14
|
* CarpetReduce: Correct weight calculationErik Schnetter2011-12-14
|
* CarpetReduce: Remove unused variable "mask"Erik Schnetter2011-12-14
|
* Merge after undoErik Schnetter2011-12-14
|\
| * Backed out changeset 1b5c07de6243Erik Schnetter2011-12-14
| |
* | CarpetReduce: Do not check for errors when enabling or disabling storageErik Schnetter2011-12-14
|/ | | | | | Do not check for errors when enabling or disabling storage for helper variables. The reported return values do not seem to conform to the specification.
* CarpetReduce: Remove unused variable "mask"Erik Schnetter2011-12-14
|
* CarpetReduce: Correct errors in calculating the weight functionErik Schnetter2011-12-14
|
* CarpetReduce: Correct error in calculating weight for restricted regionErik Schnetter2011-12-14
|
* CarpetReduce: Remove large sections of unused codeErik Schnetter2011-12-14
|
* CarpetReduce: Make helper functions min/max/etc. privateErik Schnetter2011-12-14
|
* CarpetReduce: Replace vector.at by vector.ATErik Schnetter2011-12-14
|
* CarpetReduce: Handle current time correctly when interpolating in timeErik Schnetter2011-12-14
|
* CarpetReduce: Set up mask from new dh class entriesErik Schnetter2011-12-14
| | | | | Simplify mask setup code and improve its performance by using the new dh classes and their new entries.
* CarpetReduce: Use nan handling from CarpetLibErik Schnetter2011-12-14
|
* CarpetReduce: Update commentErik Schnetter2011-12-14
|
* Re-organise time level handlingErik Schnetter2011-12-14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Store the current Cactus time (and not a fake Carpet time) in the th "time hiearchy". This removes the now redundant "leveltimes" data structure in Carpet. Add past time levels to th, so that it can store the time for past time levels instead of assuming the time step size is constant. This allows changing the time step size during evolution. Share the time hierarchy between all maps, instead of having one time hierarchy per map. Simplify the time level cycling and time stepping code used during evolution. Improve structure of the code that loops over time levels for certain schedule bins. Introduce a new Carpet variable "timelevel", similar to "reflevel". This also makes it possible to avoid time interpolation for the past time levels during regridding. The past time levels of the fine grid then remain aligned (in time) with the past time levels of the coarse grid. This is controlled by a new parameter "time_interpolation_during_regridding", which defaults to "yes" for backwards compatibility. Simplify the three time level initialisation. Instead of initialising all three time levels by taking altogether three time steps (forwards and backwards), initialise only one past time level by taking one time step backwards. The remaining time level is initialised during the first time step of the evolution, which begins by cycling time levels, which drops the non-initialised last time level anyway. Update Carpet and the mode handling correspondingly. Update the CarpetIOHDF5 checkpoint format correspondingly. Update CarpetInterp, CarpetReduce, and CarpetRegrid2 correspondingly. Update CarpetJacobi and CarpetMG correspondingly.
* CarpetReduce: Remove debugging codeErik Schnetter2011-12-14
|
* CarpetReduce: Correct multi-processor error in initialising the reduction maskErik Schnetter2011-12-14
|
* CarpetReduce: Correct handling buffer zonesErik Schnetter2011-12-14
| | | | | | Omit buffer zones when reducing. Clean up code.
* Combine CarpetLib's INSTANTIATE and Carpet's TYPECASE mechanism into aErik Schnetter2011-12-14
| | | | | | single mechanism provided by CarpetLib. Use this mechanism everywhere.
* Import CarpetErik Schnetter2011-12-14
| | | | Ignore-this: 309b4dd613f4af2b84aa5d6743fdb6b3
* Carpet/*/test/: also activate thorn LoopControl (required by CarpetReduce) ↵Thomas Radke2008-10-30
| | | | in testsuite parfiles
* CarpetReduce: Use LoopControlErik Schnetter2008-10-03
|
* CarpetReduce: Add test cases for weight on boundariesErik Schnetter2008-09-28
|
* CarpetReduce: Correct weight on staggered boundariesErik Schnetter2008-09-26
| | | | | | | Correct the weight on staggered boundaries. Correct empty domains consisting of a single internal grid point, where this grid point should have a weight of 1.
* CarpetReduce: Use C++ cast operatorsErik Schnetter2008-07-15
|
* CarpetReduce: Handle multi-patch boundaries correctlyErik Schnetter2008-07-15
| | | | | Use MultiPatch_GetBoundarySpecification to determine the boundary specifications if it is available; otherwise, use GetBoundarySpecification.
* CarpetReduce: Don't double-count grid pointsErik Schnetter2008-02-26
| | | | darcs-hash:20080226141117-dae7b-9ce593fa0c737c838211bb689937593cb32480ca.gz
* CarpetReduce: Modernise codeErik Schnetter2008-02-19
| | | | | | Declare variables where they are used for the first time. darcs-hash:20080219052703-dae7b-bed8b44832012e1abab2e92dd4b50d79506e8638.gz
* CarpetReduce: Correct error in reduction operationsErik Schnetter2007-08-28
| | | | | | | | | | | | | Make the function initialise and reduce in the reduction operators handle the count argument as well. Introduce "combine" functions into the reduction operators. These combine results from several independent reduction operations. Use the above new functions to combine reductions, correcting an error. darcs-hash:20070828044751-dae7b-02070b6aa39d25fb91481ab6a9b45cf37b10fe2f.gz
* CarpetReduce: Take weights into account when reducing in level modeErik Schnetter2007-08-27
| | | | darcs-hash:20070827201932-dae7b-2129f2816c6dcf448bf2706cf5d49cfc83e8d540.gz
* CarpetReduce: Add support for OpenMPErik Schnetter2007-08-21
| | | | darcs-hash:20070821190958-dae7b-1ed180793dd5d6482f48aa0bbf13bd5523a4b40b.gz
* CarpetReduce: Correct schedulingErik Schnetter2007-08-21
| | | | darcs-hash:20070821190829-dae7b-9efe9944356b64e8c171e6f3917f03dee4ba1deb.gz
* CarpetReduce: Implement pointwise reduction operatorsErik Schnetter2007-05-26
| | | | | | | | | | | | | | | | | | | | | | | | Implement a new set of reduction operators for Carpet. These have the same definitions as the old operators, except that they give the same weight to all grid points. The existing operators give a larger weight to coarser grid points, so that physical reduction operators work correctly, e.g. for calculating the total mass. The new operators have the same name as the old operators plus a prefix "i" (the lower case letter "eye"). For example, the new L2 norm is called "inorm2", while the old norm continues to be called "norm2". The physical (old) operators converge to a result which is independent of the grid structure. The pointwise (new) operators still converge, but to a value which depends on the grid structure. In order to make it possible to compare situations with and without symmetries, boundary points still carry a weight of 1/2 or whatever is appropriate. That means that there exists a simple relationship between the pointwise norms with and without a symmetry boundary. darcs-hash:20070526195016-dae7b-6a7e2a39bebd9c0a6377f908a4d7f446f8288191.gz
* CarpetReduce: Set up the weight function after recoveringErik Schnetter2007-05-16
| | | | | | | Set up the weight function after recovering, since the weight function is not checkpointed. darcs-hash:20070516222541-dae7b-a62fe7ade7eb4ef4a381c5710feee91321605658.gz
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-29 04:08:56 +0000