[cig-commits] r16629 - in short/3D/PyLith/trunk: . doc/userguide/tutorials/3dhex8

[willic3 at geodynamics.org]

Author: willic3
Date: 2010-05-04 16:28:04 -0700 (Tue, 04 May 2010)
New Revision: 16629

Added:
   short/3D/PyLith/trunk/Benchmarks
Modified:
   short/3D/PyLith/trunk/doc/userguide/tutorials/3dhex8/3dhex8.lyx
Log:
Added a summary of current benchmark status.
Fixed a typo in manual.



Added: short/3D/PyLith/trunk/Benchmarks
===================================================================
--- short/3D/PyLith/trunk/Benchmarks	                        (rev 0)
+++ short/3D/PyLith/trunk/Benchmarks	2010-05-04 23:28:04 UTC (rev 16629)
@@ -0,0 +1,84 @@
+#----------------------------------------------------------------------
+A list of existing benchmarks, and notes about each:
+#----------------------------------------------------------------------
+2.5D/benchmarks/savageprescott:
+This benchmark runs but runs out of iterations before converging.
+The solver/dimensionalization settings probably need updating.
+This benchmark is described in the manual but is probably out of date.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/dynamic/halfspace_strikeslip:
+This benchmark runs for both the hex8 and tet4 cases.
+There is some debugging output that we may want to remove, and we may want
+to update some PETSc settings.
+This benchmark is not described in the manual.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/dynamic/scecdynrup/tpv205:
+This benchmark was missing the output directory. It ran after that was
+added.
+This benchmark is not described in the manual.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/dynamic/scecdynrup/tpv205-2d:
+This benchmark runs for both the 100m and 200m cases.
+We may want to update some of the PETSc logging info.
+This benchmark is not described in the manual.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/trunk/quasistatic/dike-disloc:
+This benchmark needs an update of the .cfg files (doesn't work).
+This benchmark is not described in the manual.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/trunk/quasistatic/reversenog:
+This benchmark needs an update of the .cfg files (runs but has warnings).
+We also probably need to update some of the output settings (e.g., use
+time scaling in the output file names).
+This benchmark is not described in the manual.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/trunk/quasistatic/solvertest:
+This benchmark may need updating of PETSc settings. There is a warning about
+options that were not used:
+WARNING! There are options you set that were not used!
+WARNING! could be spelling mistake, etc!
+Option left: name:-sub_pc_factor_shift_nonzero no value
+Option left: name:-sub_pc_factor_shift_positive_definite value: 0
+This benchmark is not described in the manual. I don't think it is really
+a benchmark.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/trunk/quasistatic/strikeslipnog:
+This benchmark runs, but doesn't converge because it runs out of iterations.
+We probably need to adjust the solver/dimensionalization settings.
+This benchmark is described in the manual, but is probably out of date.
+
+#----------------------------------------------------------------------
+3D/PyLith/benchmarks/trunk/static/sceccvmh:
+Neither the volume or slice problems work.  The .cfg files probably need
+updating.
+This benchmark is not described in the manual and is not really a benchmark.
+
+#----------------------------------------------------------------------
+Future benchmarks?
+#----------------------------------------------------------------------
+1.  Hole in an elastic whole-space:
+This is a rough test of Drucker-Prager plasticity (current solution is for
+Mohr-Coulomb). The problem is really 2D and we would be representing it in
+3D.
+
+2.  Viscoelastic tests for a brick-shaped domain:
+These tests would work for either Maxwell or power-law. This is just a
+uniaxial strain test where we have analytical solutions for the
+time-dependent stress relaxation. The existing mesh in examples/3d/hex8
+should work fine for both.
+
+#----------------------------------------------------------------------
+Ideas for cleaning things up:
+#----------------------------------------------------------------------
+1.  Move everything int 3D/PyLith/benchmarks?
+
+2.  If something is not really a benchmark, should we have another directory
+    (3D/PyLith/tests) that allows full-scale testing with large meshes, but
+    would not need to be described in the manual?

Modified: short/3D/PyLith/trunk/doc/userguide/tutorials/3dhex8/3dhex8.lyx
===================================================================
--- short/3D/PyLith/trunk/doc/userguide/tutorials/3dhex8/3dhex8.lyx	2010-05-04 23:00:59 UTC (rev 16628)
+++ short/3D/PyLith/trunk/doc/userguide/tutorials/3dhex8/3dhex8.lyx	2010-05-04 23:28:04 UTC (rev 16629)
@@ -1,4 +1,4 @@
-#LyX 1.6.2 created this file. For more info see http://www.lyx.org/
+#LyX 1.6.5 created this file. For more info see http://www.lyx.org/
 \lyxformat 345
 \begin_document
 \begin_header
@@ -741,7 +741,7 @@
 \begin_inset Caption
 
 \begin_layout Plain Layout
-Color contours and vectors of displacement for the axial displacement example
+Color contours and vectors of displacement for the shear displacement example
  using a mesh composed of trilinear hexahedral cells generated by CUBIT.
 \begin_inset CommandInset label
 LatexCommand label