a890411 | Wolfgang Bangerth | 25 January 2015, 20:06:45 UTC | take over db3d963 Conflicts: doc/release-tasklist | 26 January 2015, 13:29:06 UTC |
8c1e1e9 | Timo Heister | 23 January 2015, 17:45:09 UTC | release task: update manual | 23 January 2015, 17:45:09 UTC |
a8eccba | Timo Heister | 23 January 2015, 17:44:08 UTC | remove broken periodic_box for now | 23 January 2015, 17:44:08 UTC |
1b5b0c5 | Timo Heister | 23 January 2015, 17:38:16 UTC | release task: remove table model | 23 January 2015, 17:38:16 UTC |
24882a3 | Timo Heister | 23 January 2015, 17:37:57 UTC | release task: update version info | 23 January 2015, 17:37:57 UTC |
16c5141 | Timo Heister | 22 January 2015, 22:15:55 UTC | Merge pull request #260 from tjhei/rename_morency_plugin Rework morency plugin - rename to "Morency and Doin" - do not parse model depth and n_compositional_fields from the ParameterHandler | 22 January 2015, 22:15:55 UTC |
395f1c6 | Timo Heister | 22 January 2015, 17:16:44 UTC | update parameters | 22 January 2015, 18:56:47 UTC |
312d63b | Timo Heister | 22 January 2015, 17:16:22 UTC | rename Morency and Doin model and section | 22 January 2015, 18:56:47 UTC |
9b05e58 | Timo Heister | 22 January 2015, 17:16:04 UTC | do not parse model depth and n_compositional_fields manually | 22 January 2015, 18:56:42 UTC |
59e1a2f | Wolfgang Bangerth | 22 January 2015, 16:16:03 UTC | Merge pull request #259 from tjhei/formatting Pre-release cleanup | 22 January 2015, 16:16:03 UTC |
4f52fdd | Timo Heister | 22 January 2015, 14:33:32 UTC | update parameters.tex | 22 January 2015, 14:33:32 UTC |
3836111 | Timo Heister | 22 January 2015, 14:31:00 UTC | rename variable | 22 January 2015, 14:31:00 UTC |
5c14b41 | Timo Heister | 22 January 2015, 14:21:45 UTC | update copyright | 22 January 2015, 14:21:45 UTC |
53d645f | Timo Heister | 22 January 2015, 14:19:45 UTC | astyle | 22 January 2015, 14:19:45 UTC |
5b73de4 | Timo Heister | 22 January 2015, 14:17:50 UTC | wrapcomments | 22 January 2015, 14:18:29 UTC |
486c413 | Timo Heister | 22 January 2015, 14:15:10 UTC | manual comment formatting | 22 January 2015, 14:18:25 UTC |
2fcde67 | Wolfgang Bangerth | 20 January 2015, 14:54:18 UTC | Merge pull request #258 from tjhei/petsc_tests add petsc mesh refinement test | 20 January 2015, 14:54:18 UTC |
327ebd4 | Timo Heister | 20 January 2015, 13:47:57 UTC | Merge pull request #256 from tjhei/depth_avg_zones change depth_average zones from 100 to 10 In each cell we place 10 quadrature points, which means 100 zones require probably 50 cells in depth direction, which typical computations do not have. | 20 January 2015, 13:47:57 UTC |
c84033c | Timo Heister | 20 January 2015, 13:21:54 UTC | add petsc mesh refinement test | 20 January 2015, 13:45:28 UTC |
925399c | Wolfgang Bangerth | 20 January 2015, 04:07:28 UTC | Merge pull request #257 from ian-r-rose/fix_petsc_calls Fix petsc calls | 20 January 2015, 04:07:28 UTC |
0e9e4c2 | ian-r-rose | 20 January 2015, 03:53:52 UTC | Workaround for is_contiguous() assertion failing in PETSc wrappers. | 20 January 2015, 03:53:52 UTC |
19b3d19 | ian-r-rose | 20 January 2015, 03:33:26 UTC | Add missing compress calls. | 20 January 2015, 03:33:26 UTC |
882fa67 | Timo Heister | 20 January 2015, 01:42:54 UTC | update tests | 20 January 2015, 02:50:33 UTC |
d75f6a6 | Timo Heister | 18 January 2015, 22:43:07 UTC | change depth_average zones from 100 to 10 In each cell we place 10 quadrature points, which means 100 zones require probably 50 cells in depth direction, which typical computations do not have. | 20 January 2015, 02:50:30 UTC |
09cd1e0 | Timo Heister | 19 January 2015, 01:41:37 UTC | Merge pull request #248 manually Change averaging in dynamic topography postprocessor from volume weighted to surface area weighted so that in the end the average of dynamic topography over the domain's surface area is zero (if Subtract mean of dynamic topography is set to true). Conflicts: tests/dynamic_topography/dynamic_topography.00000 | 19 January 2015, 01:48:37 UTC |
ac54b57 | Jacky Austermann | 16 January 2015, 22:01:54 UTC | Updating location and naming in DT postprocessor | 19 January 2015, 01:00:16 UTC |
32500e5 | Jacky Austermann | 15 January 2015, 18:47:01 UTC | Update of dynamic topography postprocessors and test | 19 January 2015, 01:00:12 UTC |
eef6299 | Timo Heister | 18 January 2015, 22:37:50 UTC | Merge pull request #242 from tjhei/pull_229 Use a manifold to generate better meshes for the shell. | 18 January 2015, 22:37:50 UTC |
f2e43e0 | Timo Heister | 16 January 2015, 20:47:03 UTC | fix test output | 17 January 2015, 04:06:57 UTC |
546de43 | Timo Heister | 12 January 2015, 01:32:19 UTC | Use higher order mapping in interior, correct manifold for shell - set straight boundary for quarter/half shell walls - only set boundary objects as needed for spherical shell - remove deal.II version check - use higher order mapping also in the interior | 17 January 2015, 04:06:57 UTC |
f050b9b | Wolfgang Bangerth | 31 October 2014, 18:37:29 UTC | Use a manifold to generate better meshes for the shell. This only works with a sufficiently new deal.II, of course. | 17 January 2015, 04:06:47 UTC |
23371c5 | Timo Heister | 16 January 2015, 19:59:56 UTC | Merge pull request #253 from tjhei/new_has_curved_el New GeometryModel::has_curved_elements() New: GeometryModel::has_curved_elements() will allow for optimizations if all boundaries are straight (and later only FlatManifold is used). | 16 January 2015, 19:59:56 UTC |
e3526ef | Timo Heister | 16 January 2015, 15:26:31 UTC | fix test output | 16 January 2015, 19:32:31 UTC |
fe3cf22 | Timo Heister | 15 January 2015, 22:29:43 UTC | Introduce GeometryModel::has_curved_elements() This allows us to switch to a Q1 mapping for box-like geometries. | 16 January 2015, 19:09:54 UTC |
a402e9e | Timo Heister | 16 January 2015, 19:08:20 UTC | Merge pull request #252 from tjhei/fix_initial_guess Fix inner solver initial guess Fixed: the inner solvers for the Stokes block preconditioner now start with a zero initial guess. This should make the solvers more robust. | 16 January 2015, 19:08:20 UTC |
c5a4c88 | Timo Heister | 16 January 2015, 15:38:05 UTC | fix test output | 16 January 2015, 17:53:31 UTC |
44716eb | Timo Heister | 16 January 2015, 13:16:25 UTC | set initial guess for inner solvers to zero | 16 January 2015, 17:48:47 UTC |
3f763d3 | Timo Heister | 16 January 2015, 16:06:46 UTC | Merge pull request #254 from bangerth/improve-error-messages Adjust a few more places where we can produce better error messages. These places were written with ExcNoConvergence in mind, but if the underlying Trilinos solver fails, we get an ExcMessage that was not printed. | 16 January 2015, 16:06:46 UTC |
9739ec6 | Wolfgang Bangerth | 16 January 2015, 15:43:11 UTC | Adjust a few more places where we can produce better error messages. These places were written with ExcNoConvergence in mind, but if the underlying Trilinos solver fails, we get an ExcMessage that was not printed. | 16 January 2015, 15:43:11 UTC |
2c1b490 | Timo Heister | 16 January 2015, 00:29:03 UTC | Merge pull request #251 from bangerth/improve-error-message Be less strict in which types of exceptions we catch. Using the change, we can also catch exceptions that do not just indicate that the solver has run out of iterations, but also exceptions that state that the underlying AztecOO solver from Trilinos has failed with this matrix. | 16 January 2015, 00:29:03 UTC |
e511a5b | Wolfgang Bangerth | 15 January 2015, 23:35:57 UTC | Be less strict in which types of exceptions we catch. Using the change, we can also catch exceptions that do not just indicate that the solver has run out of iterations, but also exceptions that state that the underlying AztecOO solver from Trilinos has failed with this matrix. | 15 January 2015, 23:38:20 UTC |
14d2520 | Jacky Austermann | 14 January 2015, 23:59:48 UTC | Change averaging in dynamic topography postprocessor from volume weighted to surface area weighted | 15 January 2015, 00:11:08 UTC |
01f19ef | Timo Heister | 13 January 2015, 19:48:26 UTC | Merge pull request #247 from tjhei/pull_234 Increase limit for number of inner iterations and reduce AMG aggregation threshold Also see #234. | 13 January 2015, 19:48:26 UTC |
36e4420 | Rene Gassmöller | 13 January 2015, 19:46:39 UTC | Merge pull request #246 from gassmoeller/check_boundary_temperature Fix tests and allow for no prescribed boundary temperature. | 13 January 2015, 19:46:39 UTC |
d7b36a8 | Rene Gassmoeller | 13 January 2015, 19:45:28 UTC | Update comment. | 13 January 2015, 19:45:28 UTC |
9d359c8 | Timo Heister | 13 January 2015, 14:52:36 UTC | fix test output | 13 January 2015, 18:02:15 UTC |
7bdad81 | Rene Gassmoeller | 13 January 2015, 17:25:38 UTC | Fix tests and allow for no prescribed boundary temperature. | 13 January 2015, 17:26:14 UTC |
ef6253b | Wolfgang Bangerth | 09 December 2014, 01:45:26 UTC | Add some commentary. | 13 January 2015, 16:38:58 UTC |
fdc9f02 | Wolfgang Bangerth | 05 December 2014, 23:24:26 UTC | Reduce the aggregation threshold for the AMG. On the Davies et al. benchmark (case 1.1) this reduces the time for the Stokes solver from 436 to 352 seconds. On the case discussed previously in this pull request, it reduces the solver time almost by a factor of 2. | 13 January 2015, 16:38:58 UTC |
1acb52a | Wolfgang Bangerth | 01 December 2014, 17:06:27 UTC | Increase limit for the number of inner iterations. This is apparently necessary for some very ill-conditioned models. Specifically, I have been running a model by Sarah Stamps with very large variations of the viscosity, and it takes more than 2000 inner iterations to converge in this place. Maybe not surprisingly, the run time profile looks rather unfamiliar: +---------------------------------------------+------------+------------+ | Total wallclock time elapsed since start | 4.37e+05s | | | | | | | Section | no. calls | wall time | % of total | +---------------------------------+-----------+------------+------------+ | Assemble Stokes system | 2 | 82.4s | 0.019% | | Assemble temperature system | 1 | 44.5s | 0.01% | | Build Stokes preconditioner | 1 | 97.3s | 0.022% | | Build temperature preconditioner| 1 | 2.77s | 0.00063% | | Solve Stokes system | 2 | 4.35e+05s | 1e+02% | | Solve temperature system | 1 | 6.15s | 0.0014% | | Initialization | 2 | 18s | 0.0041% | | Postprocessing | 1 | 1.71e+03s | 0.39% | | Setup dof systems | 1 | 60.4s | 0.014% | +---------------------------------+-----------+------------+------------+ Note that we spend 99.5% of the time in the Stokes solver. | 13 January 2015, 16:37:51 UTC |
a35a082 | Timo Heister | 13 January 2015, 15:46:46 UTC | Merge pull request #244 from tjhei/req_deal82 require deal.II 8.2 and remove old checks | 13 January 2015, 15:46:46 UTC |
ec114d6 | Timo Heister | 13 January 2015, 15:45:40 UTC | add changelog entry | 13 January 2015, 15:45:40 UTC |
511d40d | Timo Heister | 13 January 2015, 15:11:49 UTC | require deal.II 8.2 and remove old checks | 13 January 2015, 15:11:49 UTC |
7eb5dd7 | Wolfgang Bangerth | 13 January 2015, 13:45:46 UTC | Merge pull request #243 from gassmoeller/update_composition_boundaries Update composition boundaries | 13 January 2015, 13:45:46 UTC |
cbf1eae | Rene Gassmoeller | 13 January 2015, 11:15:11 UTC | Unify temperature and composition boundary conditions. | 13 January 2015, 11:15:11 UTC |
9d9ceee | Rene Gassmoeller | 13 January 2015, 00:47:03 UTC | Add update function for boundary composition. | 13 January 2015, 00:47:03 UTC |
1a11b0a | Timo Heister | 10 January 2015, 00:21:42 UTC | Merge pull request #241 from tjhei/test_output fix test output | 10 January 2015, 00:21:42 UTC |
2c49149 | Timo Heister | 31 December 2014, 21:57:04 UTC | correct test output Some of the output changed with the switch to deal 8.2 | 09 January 2015, 16:39:18 UTC |
fef0a8c | Timo Heister | 19 December 2014, 21:49:37 UTC | Merge pull request #239 from tjhei/fix_vis_output_interval fix visualization output interval This fixes the test composition-passive-tracers which happens to reach a time step that is an exact multiple of the output interval. Bug was introduced in 8e658c7. | 19 December 2014, 21:49:37 UTC |
4c5e243 | Timo Heister | 18 December 2014, 08:42:39 UTC | fix visualization output interval This fixes the test composition-passive-tracers which happens to reach a time step that is an exact multiple of the output interval. Bug was introduced in 8e658c7b94f. | 19 December 2014, 17:55:59 UTC |
1a6b773 | Timo Heister | 18 December 2014, 07:23:50 UTC | Merge PR #235 'Changed the box geometry to allow you to specify an origin' Conflicts: doc/modules/changes.h | 18 December 2014, 07:23:50 UTC |
c1465eb | Wolfgang Bangerth | 11 December 2014, 21:52:10 UTC | Merge pull request #238 from jperryhouts/fix_minor_typo Fix minor typo in a comment. | 11 December 2014, 21:52:10 UTC |
94a8860 | Jonathan Perry-Houts | 11 December 2014, 21:28:58 UTC | Fix minor typo in a comment. | 11 December 2014, 21:28:58 UTC |
ffc256c | Wolfgang Bangerth | 10 December 2014, 21:59:52 UTC | Merge pull request #158 from jperryhouts/morency_doin_material Added material model based on rheology in Morency and Doin, 2004 | 10 December 2014, 21:59:52 UTC |
2d0c744 | Jonathan Perry-Houts | 10 July 2014, 22:25:06 UTC | Added material model based on rheology in Morency and Doin, 2004 | 10 December 2014, 19:29:15 UTC |
7b5ce1a | Ryan Grove | 09 December 2014, 19:11:10 UTC | Added test | 09 December 2014, 19:11:10 UTC |
ff44215 | Ryan Grove | 09 December 2014, 19:10:38 UTC | Fixed initial conditions for box with origin | 09 December 2014, 19:10:38 UTC |
cf064c4 | Ryan Grove | 09 December 2014, 16:09:55 UTC | Addressed issues and added to changelog Fixed labels to include units with proper grammar and also added a blurb to the changelog. | 09 December 2014, 16:09:55 UTC |
8e658c7 | Rene Gassmöller | 04 December 2014, 17:04:54 UTC | Merge pull request #236 from gassmoeller/visualization_fix Always create a visualization output after restart if interval is set to 0 | 04 December 2014, 17:04:54 UTC |
8b38e5e | Rene Gassmoeller | 02 December 2014, 14:25:33 UTC | Handle change of output_interval during restart. | 03 December 2014, 16:31:53 UTC |
66c0a9b | Timo Heister | 01 December 2014, 20:54:46 UTC | Merge pull request #232 from jdannberg/iterated_IMPES_test add a test for the iterated IMPES solver | 01 December 2014, 20:54:46 UTC |
2ce551b | Ryan Grove | 01 December 2014, 18:37:28 UTC | Changed the box geometry to allow you to specify an origin. | 01 December 2014, 18:44:10 UTC |
86d8c86 | Juliane Dannberg | 17 November 2014, 19:43:06 UTC | add a test for the iterated IMPES solver | 17 November 2014, 19:43:06 UTC |
822ab17 | Timo Heister | 14 November 2014, 22:12:14 UTC | Merge pull request #231 from bangerth/doc-update Update comments in a couple of places. | 14 November 2014, 22:12:14 UTC |
afc56e9 | Wolfgang Bangerth | 14 November 2014, 20:35:02 UTC | Update comments in a couple of places. | 14 November 2014, 20:35:02 UTC |
4812dd6 | Timo Heister | 11 November 2014, 20:49:49 UTC | Merge pull request #230 from jdannberg/iterated_IMPES_fix fix the residual of the iterated IMPES solver | 11 November 2014, 20:49:49 UTC |
12895d0 | Juliane Dannberg | 11 November 2014, 19:45:24 UTC | put comment back in the right place | 11 November 2014, 19:45:24 UTC |
9a9fdd2 | Juliane Dannberg | 11 November 2014, 15:29:02 UTC | fix the residuum of the iterated IMPES solver | 11 November 2014, 15:29:02 UTC |
aed4d3e | Rene Gassmöller | 29 October 2014, 18:30:00 UTC | Merge pull request #228 from gassmoeller/gplate_fix Gplate fix | 29 October 2014, 18:30:00 UTC |
9bf63b5 | Rene Gassmoeller | 29 October 2014, 10:27:33 UTC | Add test for gplates-3d. | 29 October 2014, 14:57:32 UTC |
aad3999 | Timo Heister | 29 October 2014, 13:09:24 UTC | Merge pull request #227 from SiqiZhang/gplate_fix Fix a bug in gplate plugin: While gplate uses zero longitude at all grid points at poles, we assume the longitude at those points vary from 0 to 360. This results in a wrong velocity to be interpolated near the poles (the velocity between poles to the first latitude grid points would be wrong, however the velocity right at the pole will be OK). | 29 October 2014, 13:09:24 UTC |
c534a6e | Siqi Zhang | 28 October 2014, 02:26:07 UTC | Add comment | 28 October 2014, 02:26:07 UTC |
64af8ca | Siqi Zhang | 27 October 2014, 00:15:32 UTC | Fix a bug that gives wrong velocity near the poles. | 28 October 2014, 00:35:31 UTC |
a4d34fe | Timo Heister | 10 October 2014, 20:56:48 UTC | Merge pull request #225 from bangerth/fix-previous-commit Fix compilation error introduced in 6a39ed4. | 10 October 2014, 20:56:48 UTC |
a480534 | Wolfgang Bangerth | 10 October 2014, 20:12:21 UTC | Fix compilation error introduced in 6a39ed4. | 10 October 2014, 20:12:21 UTC |
7f59983 | Wolfgang Bangerth | 10 October 2014, 19:33:42 UTC | Merge pull request #223 from tjhei/fix_deal_81_compile fix c8a76830 with deal 8.1 | 10 October 2014, 19:33:42 UTC |
6a39ed4 | Timo Heister | 10 October 2014, 19:23:05 UTC | fix c8a76830 with deal 8.1 | 10 October 2014, 19:23:05 UTC |
2e8a68d | Wolfgang Bangerth | 10 October 2014, 19:08:01 UTC | Merge pull request #222 from tjhei/mac_link_h_fix2 guard symbol usage of link.h only when we have it | 10 October 2014, 19:08:01 UTC |
6748db2 | Timo Heister | 10 October 2014, 17:40:04 UTC | guard symbol usage of link.h only when we have it This should fix compilation on mac os, where we have no link.h. Note that we are not calling get_names_of_shared_libs() in this case anyway. | 10 October 2014, 17:40:04 UTC |
2758b7d | Wolfgang Bangerth | 02 October 2014, 22:33:14 UTC | Merge pull request #221 from tjhei/fix_tests fix test output | 02 October 2014, 22:33:14 UTC |
9e61258 | Timo Heister | 02 October 2014, 20:12:58 UTC | Merge pull request #220 from bangerth/verify-link.h-existence Make sure <link.h> exists before using it. | 02 October 2014, 20:12:58 UTC |
ee0d59b | Timo Heister | 02 October 2014, 19:46:12 UTC | fix test output some differences were introduced in the multicomponent* tests and the other tests changed due to some configuration change. | 02 October 2014, 19:46:12 UTC |
edfb97a | Wolfgang Bangerth | 02 October 2014, 15:54:05 UTC | Merge pull request #219 from ian-r-rose/fix_typo Fix bug in multicomponent | 02 October 2014, 15:54:05 UTC |
e4cd7c5 | Wolfgang Bangerth | 02 October 2014, 12:10:17 UTC | Make sure <link.h> exists before using it. | 02 October 2014, 12:10:17 UTC |
c8a7683 | Timo Heister | 02 October 2014, 11:35:12 UTC | Merge pull request #217 from bangerth/accelerate-sparsity-pattern Use a trick by Martin Kronbichler to accelerate matrix operations in parallel | 02 October 2014, 11:35:12 UTC |
132bc47 | ian-r-rose | 01 October 2014, 21:33:10 UTC | Add a check to avoid this bug. | 01 October 2014, 21:33:10 UTC |
6e51f66 | ian-r-rose | 01 October 2014, 21:11:51 UTC | These should be lowercase. | 01 October 2014, 21:11:51 UTC |
b27e3e0 | Wolfgang Bangerth | 30 September 2014, 19:52:42 UTC | Use a trick by Martin Kronbichler to accelerate matrix operations in parallel. This tells deal.II to use a structure on the side to facilitate faster writing into matrices when doing computations in parallel. | 30 September 2014, 19:52:42 UTC |
9b2e432 | Timo Heister | 19 September 2014, 18:08:09 UTC | Merge pull request #216 from jdannberg/melt_fraction_fix adjust a default value to account for the use of a different temperature in the original source of the melting model. I just noticed that in the Sobolev et al (2011) paper they use the potential temperature instead of the absolute temperature. This means they do not consider the adiabatic temperature change and their melting temperature is lower compared to the one that should be used in a model with the "real" temperature. | 19 September 2014, 18:08:09 UTC |
623ff74 | Juliane Dannberg | 19 September 2014, 16:52:25 UTC | adjust a default value to account for the use of a different temperature in the original source of the melting model | 19 September 2014, 16:52:25 UTC |