[cig-commits] r21188 - seismo/3D/SPECFEM3D/trunk

[ampuero at geodynamics.org]

Author: ampuero
Date: 2012-12-20 18:02:19 -0800 (Thu, 20 Dec 2012)
New Revision: 21188

Modified:
   seismo/3D/SPECFEM3D/trunk/README_SPECFEM3D_FAULT
   seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt
Log:
added note about consolidating ids in cubit; added faults to-do list

Modified: seismo/3D/SPECFEM3D/trunk/README_SPECFEM3D_FAULT
===================================================================
--- seismo/3D/SPECFEM3D/trunk/README_SPECFEM3D_FAULT	2012-12-20 09:36:26 UTC (rev 21187)
+++ seismo/3D/SPECFEM3D/trunk/README_SPECFEM3D_FAULT	2012-12-21 02:02:19 UTC (rev 21188)
@@ -149,8 +149,13 @@
 which looks like the example above. Variables can be defined and manipulated there using the 
 APREPRO language built in Cubit.
 
+Note: you should avoid gaps in the list of indices of mesh objects 
+with the following Cubit command:
+  compress ids hex face edge node')
+(otherwise you will get a segmentation fault during domain decomposition)
 
 
+
 CUBIT-PYTHON SCRIPTS FOR FAULTS
 -------------------------------
 

Modified: seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt
===================================================================
--- seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt	2012-12-20 09:36:26 UTC (rev 21187)
+++ seismo/3D/SPECFEM3D/trunk/todo_list_please_dont_remove.txt	2012-12-21 02:02:19 UTC (rev 21188)
@@ -446,6 +446,35 @@
 For some of Paul Cristini's 2D simulations for instance the current ASCII file become really huge
 
 
+------------------------------------------------
+dynamic/kinematic faults:
+------------------------------------------------
+
++ fault_solver_common:
+	- make ordered version of subroutine assemble_MPI_vector_ext_mesh, and use it in subroutine initialize_fault 
+
++ fault_solver_dynamic:
+	- many hard-coded ad hoc features need to be set instead as user options
+
++ rate-and-state friction:
+	- make it a user-friendly option (currently a flag in fault_solver)
+
++ cubit interface:
+	- consolidate python scripts (eliminate redundancy)
+
++ verify consistency of fault edge nodes in kinematic solver: 
+  currently, these nodes are not split but they are included in the fault database
+
++ simplify mesh generation with faults:
+ 	- eliminate fault edge nodes? 
+          Currently fault edge nodes must be defined as closed and non-split in CUBIT,
+          SESAME asigns a single ibulk, but it adds it to the fault database.
+	- develop a recipe for mesh coarsening in CUBIT based on half "cubed sphere" mesh
+        - test tet-to-hex strategy
+
++ add snapshot outputs to kinematic solver (e.g. to export fault stresses for adjoint source inversion)
+
+
 Older to-do lists of SPECFEM2D, SPECFEM3D and SPECFEM3D_GLOBE (some items are probably still useful to implement):
 ------------------------------------------------------------------------------------------------------------------