https://github.com/geodynamics/citcoms

sort by:
Revision Author Date Message Commit Date
70e5d7d Minor fix on node counting 05 November 2008, 00:19:05 UTC
00db2ff Added a missing import 04 November 2008, 22:50:02 UTC
3c39d9c zslice.py will slice opt files if present 04 November 2008, 21:02:08 UTC
629713a Removed input parameters "mgunitx", "mgunity" and "mgunitz". Their values are inferred from other parameters. Also, the code checks that "levels" > 1 when multigrid solver is used. 04 November 2008, 21:01:26 UTC
dc54393 Fixed a bug in outputting comp_el 04 November 2008, 21:00:01 UTC
4b191bb Fixed two typos, velocity norm function was called where pressure norm should have been called. This leads to core dumps, all seems to be working now. 01 November 2008, 18:37:20 UTC
f8e52fc Reporting correct nno and nel (number of nodes and elements) 31 October 2008, 20:18:13 UTC
ef7299c (for r13196) Fixing convergence criterion for Stokes solver. * New functions global_v_norm2(), global_p_norm2(), global_div_norm2() compute the L2 norm of velocity, pressure, and divergence. * The L2 norm of the vector f is defined as: \sum_{all elements} ( \int_{element} (f*f dA) ) / A where A is the volume of the domain. * The convergence criterion is controlled by parameter "accuracy" under "[CitcomS.solver.vsolver]". The iteration stops if (norm(div(rho*V))/norm(V) < accuracy) or ( (norm(dV)/norm(V) < accuracy) and (norm(dP)/norm(P) < accuracy) for the last two iterations ) * The default value of "accuracy" becomes 1e-4. * The input parameters "tole_compressibility" and "relative_err_accuracy" under "[CitcomS.solver.vsolver]" are gone. * The mass matrix E->MASS becomes double precision (was single precision). 29 October 2008, 23:22:10 UTC
2ba507e Fixed two bugs in vtk velocity output 29 October 2008, 23:17:11 UTC
f52d984 Initial test implementation of netcdf grd based assignment of designated Euler vectors based on a code grd. 23 October 2008, 23:35:27 UTC
f2bb425 multi-component chemical viscosity 04 September 2008, 22:46:39 UTC
fed8acd Fixed issue 158: incorrect E->control.fi_max with coor=1 04 September 2008, 20:18:53 UTC
e2213ae Some basic VTK parallel XML output * output in vts/pvts Structured VTK XML format * only these fields are available: coordinate, temperature, velocity and viscosity * coordinate and velocity are in Cartesian coordinates Some notes for future tasks: * adding other node- or element-based fields (pressure, stress, comp_nd, comp_el) is easy * tracer output should be in vtp format * surf, botm, and horiz_avg will be in vts format * geoid will still be as an ascii table 04 September 2008, 00:20:50 UTC
46cfb7a forgot to checkin this change in r12799 04 September 2008, 00:12:24 UTC
6ef0cc0 pyre input for "self_gravitation" 04 September 2008, 00:11:36 UTC
9eff47c self-gravitation geoid stuff * compute geoid due to internal buoyancy at the bottom * get rid of stress arrays * seperating topo effect on surface geoid into two part: surface topo and CMB topo. 04 September 2008, 00:10:51 UTC
867f7be E->viscosity.zbase_layer is not init'd in pyre version. This fixes issue 155. 24 July 2008, 18:42:45 UTC
972d8eb Both Eh and I were right and now the z_layer check should be finally fixed. 24 July 2008, 18:21:37 UTC
ccc8ab9 Fixed typo that provided assignment of zbase_layer thanks to Eh. 24 July 2008, 18:13:22 UTC
cc2e6d9 Added rheology option eight. The other merges are surprising. 24 July 2008, 01:00:16 UTC
55333fb Moved the stress computation into stress output subroutine, for consistency with the style of the Output...c files. 30 June 2008, 15:49:56 UTC
871fd84 Made CBF topography method and call to E->solver.parallel_communication_routs_s contingent on a "use_cbf_topo" parameter which is for now, by default, off, until the potential remaining memory bug (?) is tracked down. Made sure the stress tensor is computed before stress output, as CBF topo does not require the computation of the stress tensor. Added a z_layer input flag for zbase_layer other than the four layers used for the control of phase boundary depth. (Default is backward compatible.) 30 June 2008, 02:12:56 UTC
eb573f1 Check in of proper version for elliptical net rotation correction. 27 June 2008, 00:53:30 UTC
9f7396f Reordered loop in get_buoyancy to speed up array access. Fixed omission of one elliptical correction, added a flag. 26 June 2008, 23:56:32 UTC
fa90e33 Replaced several trig function evaluations in Size_does_matter.c witht he precomputed SinCos arrays for speedup. Experimental ellipticity implementation has been improved somewhat, and made accessible only when -DALLOW_ELLIPTICAL is set. This was necessary to not slow the spherical version of the code down with if statements. 26 June 2008, 23:04:06 UTC
cc26d94 Modified net rotation computation for ellipticity. Minor additions (prototypes) to make the code compile smoothly under PGI. 25 June 2008, 01:13:32 UTC
6ada82f Modified routs_s for lines. Added some debugging output and updated comments. 24 June 2008, 21:18:37 UTC
a0ca2a8 Only top/bottom processors need to communicate 24 June 2008, 21:11:18 UTC
30e6081 Assignment of velocity boundary condition only called once at the top multigrid level during boundary condition assignment. Experimental implementation of mixed velocity/free slip boundary condition based on netcdf grids, in Ggrd_handling. 21 June 2008, 22:28:51 UTC
e6aaa4d Enable CBF topography 20 June 2008, 23:01:28 UTC
9cab369 Added back get_CBF_topo(), from Shijie Zhong, with some minor modifications. Minor modifications: * some double arrays -> float arrays * calling exchange_snode_d() -> calling *_exchange_snode_f() 20 June 2008, 22:51:56 UTC
1f78588 Resurrected full_parallel_communication_routs_s(), which is needed by CBF topo. Renamed exchange_snode_f() to full_exchange_snode_f() and regional_exchange_snode_f(). 20 June 2008, 22:50:51 UTC
b0228c2 Added get_global_1d_shape_fn_L() from Shijie Zhong 20 June 2008, 22:42:30 UTC
d98cafb Compute E->L and E->Lx 1d shape functions (1d is a misnomer, they are 2d actually). 20 June 2008, 22:41:57 UTC
2e17e1d The function declaration was put at a wrong place. 20 June 2008, 19:08:03 UTC
bf432bb Change back the order of input arguments to match the manual. 19 June 2008, 22:20:07 UTC
fff43af The comment in rheol=4 doesn't match the code and is removed. Added option rheol=7, with a complete temperature- and pressure-dependent Arrhenius law. This fixed issue152. 19 June 2008, 21:43:25 UTC
c3a51e5 Added debugging (but disabled) function for spherical harmonics expansion. 19 June 2008, 21:42:27 UTC
ab74087 Send geoid arrays to vertical columns of processors. After this fix, the geoid output is the same for nprocz=1 and nprocz>1. 19 June 2008, 01:05:12 UTC
08ab04b Compute E->surf_det (top surface area, ie. surface jacobian, of an element) on a unit sphere. Use E->surf_det, instead of the surface jacobian returned by get_global_1d_shape_fn(), to expand spherical harmonics. The surface jacobian returned by get_global_1d_shape_fn() is defined at the top radius of the current mesh. When nprocz>1, the top radius is not the same as the outer radius of the sphere for processors. As a result, the spherical harmonics expansion gave incorrect result when nprocz>1 in the previous revisions. 19 June 2008, 01:04:06 UTC
d3025a8 Speed up of computation by precomputing constants. 18 June 2008, 22:35:04 UTC
2d8894e Compute geoid with self gravitation. Code contributed by Shijie Zhong. 17 June 2008, 21:23:45 UTC
de569a3 Added a few more comments and made a few loops more efficient. 17 June 2008, 20:31:30 UTC
67ed449 Added a few comments on Rayleigh number netcdf control. 14 June 2008, 00:00:58 UTC
9735d41 update NEWS item 13 June 2008, 23:36:53 UTC
d5dc9d3 Replaced MPI_COMM_WORLD by E->parallel.world. MPI_COMM_WORLD is the global communicator. E->parallel.world is the communicator for the current solver. These two are different in a coupled run. Before this fix, full-regional coupling will dead lock. 13 June 2008, 23:35:38 UTC
adc7607 Added some comments 13 June 2008, 23:34:08 UTC
dbd7942 To be consistent with C version, disable checkpoint point output at 0th step if restarting. 13 June 2008, 23:33:29 UTC
35878a7 Correctly counting the global number of equations. This will slightly affect the results of the stokes solver. This avoid double counting the overlapping nodes between processors and the global norm of velocity vector. 13 June 2008, 23:32:21 UTC
e5d4428 Using firewall channel, since error channel doesn't stop the code. 13 June 2008, 23:25:41 UTC
94e1195 Decalred variables as double to avoid float-double conversion later in the loop. 13 June 2008, 23:24:32 UTC
b118cc2 update the help message 13 June 2008, 19:14:12 UTC
d2eccfa Added tentative implementation of a rotating, elliptical Earth. Made restart for constant viscosity cases possible. Added alternative geoid computation place holder for self-gravitating mantle, to be filled out by Shijie's old code. 12 June 2008, 21:22:41 UTC
6f05c7a Fixed typo in memory allocation procedure that led to core dump. 30 May 2008, 18:42:57 UTC
de1b3e0 fixed typo 12 May 2008, 21:22:42 UTC
1fbd9ec Just a place holder for the next release 12 May 2008, 21:16:55 UTC
9a0fcce Bug Fix 05 May 2008, 19:17:31 UTC
8ffc5f0 Add python binding to update top surface temperatures 01 May 2008, 22:59:14 UTC
e7216e6 More additions to allow top surface temperatures to be read in a each time step. Should read in every time step for C only version. Additional changes for Pyre version to follow; will test more once these changes are made. Small changes to viscosity structures when reading in materials for each element. More documentation to follow. 01 May 2008, 22:56:12 UTC
e137764 Inputing tbcs_file & temperature_boundary_file 01 May 2008, 00:51:25 UTC
7499aea Start to add new features for reading surface temperature for all top surface node points Added new variables to the CONTROL structure vbcs_file & temperature_boundary_fi le 30 April 2008, 23:11:18 UTC
1d4d90a GMT/Netcdf grd input can now deal with velocity boundary conditions, material dependence, and local Rayleigh number in surface layers. 06 April 2008, 21:58:45 UTC
1866b6f Close trace_log files when finalizing 05 March 2008, 21:15:02 UTC
adc1f2d Clean up unused arrays and functions 05 March 2008, 21:07:02 UTC
b33dc3c Write exclusive tracer time every 20 steps 05 March 2008, 21:06:11 UTC
331291e Revert r11328 05 March 2008, 21:04:21 UTC
1709362 Write exclusive tracer time every 20 steps 05 March 2008, 20:50:31 UTC
37cc271 Removed artificial tolerance of empty elements. This greatly decreases the minimal number of tracers per element required for a given mesh. This fixed issue140. 28 February 2008, 01:03:50 UTC
cc3e43d Minor tweak 28 February 2008, 01:00:54 UTC
e15117c Built surface velocity BC support via Netcdf grd files back in. This needs more testing, but it would be good to have the framework in place before new additions are made. Also, age control is not yet implemented, but will be soon. 27 February 2008, 07:42:44 UTC
332d337 Removed unnecessary call to construct_mat_group() 21 February 2008, 01:25:34 UTC
c42f6fd Finished interpolating fields onto tracers 21 February 2008, 01:03:43 UTC
222c941 Renamed fix_phi() to fix_angle() 21 February 2008, 01:02:41 UTC
2bd8f42 Fixed issue139. Bug for tracers near South pole. 21 February 2008, 01:01:09 UTC
c0e2815 Get new tracers in post processing stage 21 February 2008, 01:00:21 UTC
346ba4e Read multiple extraq 21 February 2008, 00:56:29 UTC
5cc21b8 Write averaged composition in horiz_avg output. The HDF5 part of output is not implemented... 21 February 2008, 00:55:41 UTC
892dd0d Removed horizontal average from output_interpolated_fields() 21 February 2008, 00:54:25 UTC
89dc1c4 Compute horizontal average for composition fields 21 February 2008, 00:53:35 UTC
a9e6ba7 Modified Citcom.c such that E->mat gets assigned based on element depth (as was default before), even if mat_control is not selected. Added (back in?) ggrd netcdf type VIP prefactor material control via ggrd_mat_control flags (as documented in Instructions.c) Renamed some of the ggrd stuff to make it more modular. 17 February 2008, 00:14:06 UTC
797ffec Interpolate fields in post_processing() * The parameter "itracer_interpolate_fields" is the same as "model_type" in the AVM code. * If itracer_interpolate_fields is 1, 2, or 3, it interpolates temperature and compositions. The horizonal average is not removed yet. 09 February 2008, 00:14:35 UTC
0262f7b Refactored the functions. * get shape functions on tracer location * interpolate data using acquired shape functions 08 February 2008, 23:58:57 UTC
a0c621f Added binding for post_processing(). Exit program after post processing and saving results. 08 February 2008, 23:56:39 UTC
8688232 * Using a different reference point for uv space. This fixed a nasty bug when some cap extends over ~pi/2. * A bunch of debugging output (disabled) * Disable more debugging output 08 February 2008, 23:54:36 UTC
aed8cda Added new parameter: itracer_interpolate_fields When it is 0 (default), nothing is done. Otherwise, the convection fields are interpolated onto the location of tracers. 08 February 2008, 23:50:30 UTC
5b88643 Reverted part of unwannted change 04 February 2008, 23:23:21 UTC
f3b65a8 Fixed potential memory leak 04 February 2008, 23:21:19 UTC
7455cf7 Fixed bugs in managing reference counts and potential double-free 04 February 2008, 23:13:23 UTC
61854c2 Fixed a regression reported by Lydia Dicaprio. The initial temperature should be constructured in this order: 1. if tic_method=-1, read from velo files. 2. if lith_age=1, read/constructed from lith_age files. 3. otherwise, call {full,regional}_construct_tic_from_input() 04 February 2008, 20:40:45 UTC
8eda467 Renamed functions 22 January 2008, 19:52:09 UTC
2c8211b Computes the derivatives of shape functions and jacobians only once and stores them for later use. This reduces the total cpu time of cookbook8 by 13%. - get_global_shape_fn_sph() computes and stores the derivatives of shape functions and jacobians. - Moved get_global_shape_fn() to Obsolete.c - Calling construct_shape_function_derivatives() in initial_mesh_solver_setup(). This function call get_global_shape_fn_sph() for each element. - get_rtf_at_vpts() and get_rtf_at_ppts() are for coord. transformation matrix at vpts and ppts respectively - Passed arguments by references in PG solvers instead of by values to avoid copying. 22 January 2008, 19:38:43 UTC
64dc706 Changed GNOMONIC to CITCOM_GNOMONIC to avoid namespace conflict with GMT gmt.h 17 January 2008, 02:06:25 UTC
07ddb1c Forgot to checkin this file 16 January 2008, 20:53:14 UTC
97b3adf Added timing information for tracer module. The timing is reported in *.tracer_log.* files. 16 January 2008, 20:34:22 UTC
077bc61 A more memory efficient way to find shape function for tracers. - Migrated E->tracer.UV to E->gnomonic. The reference point is the 1st node of the local mesh. The 0th element of E->gnomonic stores the sine of cosine of reference theta. - Convertd shape_coefs array from size of nel to size of snel to save memory. 16 January 2008, 20:30:09 UTC
1637c6f Added Luis Armendariz to the author list 15 January 2008, 23:36:54 UTC
7d95bab Simplified loop over levels 15 January 2008, 23:11:38 UTC
a12771a Reverted previous checkin 15 January 2008, 23:05:58 UTC
e30ef29 Simplified loop over levels 15 January 2008, 23:01:26 UTC
6febbb9 Forgot to rename variables 15 January 2008, 00:29:40 UTC
back to top