Merge Pull Request #4820 from gsjaardema/Trilinos/april-seacas-snapshot

Automatically Merged using Trilinos Pull Request AutoTester
PR Title: Automatic snapshot commit from seacas at ee0f81875
PR Author: gsjaardema

packages/seacas/IossProperties.md

@@ -2,25 +2,25 @@
 
   Property | Value    | Description
  ----------|----------|------------
- LOGGING   | on/[off] | enable/disable logging of field input/output
- LOWER\_CASE\_VARIABLE\_NAMES | [on]/off | Convert all variable names on database to lowercase; replace ' ' with '_'
+ LOGGING   | on/\[off] | enable/disable logging of field input/output
+ LOWER\_CASE\_VARIABLE\_NAMES | \[on]/off | Convert all variable names on database to lowercase; replace ' ' with '_'
  VARIABLE\_NAME\_CASE | upper/lower | Convert all variable names on output database to upper or lower case
- USE\_GENERIC\_CANONICAL\_NAMES | on/[off]  | use `block_{id}` as canonical name of an element block instead of the name (if any) stored on the database. The database name will be an alias.
- ENABLE\_FIELD\_RECOGNITION | [on]/off | try to combine scalar fields with common basename and recognized suffix into vector, tensor, ...
- FIELD\_SUFFIX\_SEPARATOR  | character ['_'] | use this suffix as separtor between field basename and suffices when recognizing fields
- MINIMIZE\_OPEN\_FILES | on/[off] | If on, then close file after each timestep and then reopen on next output
- CYCLE\_COUNT | {int}[infinite] | See `OVERLAY_COUNT` description below
- OVERLAY\_COUNT | {int}[0] | Output `OVERLAY_COUNT` steps to database on top of each other.  `DB_STEP = (((IOSS_STEP-1) / (OVERLAY_COUNT+1)) % CYCLE_COUNT) +1`
+ USE\_GENERIC\_CANONICAL\_NAMES | on/\[off]  | use `block_{id}` as canonical name of an element block instead of the name (if any) stored on the database. The database name will be an alias.
+ ENABLE\_FIELD\_RECOGNITION | \[on]/off | try to combine scalar fields with common basename and recognized suffix into vector, tensor, ...
+ FIELD\_SUFFIX\_SEPARATOR  | character \['_'] | use this suffix as separtor between field basename and suffices when recognizing fields
+ MINIMIZE\_OPEN\_FILES | on/\[off] | If on, then close file after each timestep and then reopen on next output
+ CYCLE\_COUNT | {int}\[infinite] | See `OVERLAY_COUNT` description below
+ OVERLAY\_COUNT | {int}\[0] | Output `OVERLAY_COUNT` steps to database on top of each other.  `DB_STEP = (((IOSS_STEP-1) / (OVERLAY_COUNT+1)) % CYCLE_COUNT) +1`
 ## Auto-Decomposition-Related Properties
 
  Property        | Value  | Description
 -----------------|--------|-----------------------------------------------------------
 MODEL\_DECOMPOSITION\_METHOD | {method} | Decompose a DB with type `MODEL` using `method`
 RESTART\_DECOMPOSITION\_METHOD | {method} | Decompose a DB with type `RESTART_IN` using `method`
 DECOMPOSITION\_METHOD | {method} | Decompose all input DB using `method`
-PARALLEL\_CONSISTENCY | [on]/off | On if the client will call Ioss functions consistently on all processors. If off, then the auto-decomp and auto-join cannot be used.
-RETAIN\_FREE\_NODES | [on]/off | In auto-decomp, will nodes not connected to any elements be retained.
-LOAD\_BALANCE\_THRESHOLD | {real} [1.4] | CGNS-Structured only -- Load imbalance permitted Load on Proc / Avg Load
+PARALLEL\_CONSISTENCY | \[on]/off | On if the client will call Ioss functions consistently on all processors. If off, then the auto-decomp and auto-join cannot be used.
+RETAIN\_FREE\_NODES | \[on]/off | In auto-decomp, will nodes not connected to any elements be retained.
+LOAD\_BALANCE\_THRESHOLD | {real} \[1.4] | CGNS-Structured only -- Load imbalance permitted Load on Proc / Avg Load
 
 ### Valid values for Decomposition Method
 
@@ -41,28 +41,28 @@ external   | Files are decomposed externally into a file-per-processor in a para
 
  Property        | Value  | Description
 -----------------|--------|-----------------------------------------------------------
-COMPOSE\_RESTART  | on/[off] |
-COMPOSE\_RESULTS  | on/[off] |
+COMPOSE\_RESTART  | on/\[off] |
+COMPOSE\_RESULTS  | on/\[off] |
 PARALLEL\_IO\_MODE | netcdf4, hdf5, pnetcdf | mpiio and mpiposix are deprecated hdf5=netcdf4
 
 ## Properties Related to byte size of reals and integers
 
  Property              | Value  | Description
 -----------------------|--------|-----------------------------------------------------------
- INTEGER\_SIZE\_DB       | [4] / 8 | byte size of integers stored on the database.
- INTEGER\_SIZE\_API      | [4] / 8 | byte size of integers used in api functions.
- REAL\_SIZE\_DB          | 4 / [8] | byte size of floating point stored on the database.
- REAL\_SIZE\_API         | 4 / [8] | byte size of floating point used in api functions.
+ INTEGER\_SIZE\_DB       | \[4] / 8 | byte size of integers stored on the database.
+ INTEGER\_SIZE\_API      | \[4] / 8 | byte size of integers used in api functions.
+ REAL\_SIZE\_DB          | 4 / \[8] | byte size of floating point stored on the database.
+ REAL\_SIZE\_API         | 4 / \[8] | byte size of floating point used in api functions.
 
 ## Properties related to underlying file type (exodus only)
 
  Property              | Value  | Description
 -----------------------|--------|-----------------------------------------------------------
-  FILE\_TYPE            | [netcdf], netcdf4, netcdf-4, hdf5 |
- COMPRESSION\_LEVEL     | [0]-9    | In the range [0..9]. A value of 0 indicates no compression, will automatically set `file_type=netcdf4`, recommend <=4
- COMPRESSION\_SHUFFLE   | on/[off] |to enable/disable hdf5's shuffle compression algorithm.
- MAXIMUM\_NAME\_LENGTH   | [32]     | Maximum length of names that will be returned/passed via api call.
- APPEND\_OUTPUT         | on/[off] | Append output to end of existing output database
+  FILE\_TYPE            | \[netcdf], netcdf4, netcdf-4, hdf5 |
+ COMPRESSION\_LEVEL     | \[0]-9    | In the range \[0..9]. A value of 0 indicates no compression, will automatically set `file_type=netcdf4`, recommend <=4
+ COMPRESSION\_SHUFFLE   | on/\[off] |to enable/disable hdf5's shuffle compression algorithm.
+ MAXIMUM\_NAME\_LENGTH   | \[32]     | Maximum length of names that will be returned/passed via api call.
+ APPEND\_OUTPUT         | on/\[off] | Append output to end of existing output database
  APPEND\_OUTPUT\_AFTER\_STEP | {step}| Max step to read from an input db or a db being appended to (typically used with APPEND\_OUTPUT)
  APPEND\_OUTPUT\_AFTER\_TIME | {time}| Max time to read from an input db or a db being appended to (typically used with APPEND\_OUTPUT)
 
@@ -71,29 +71,28 @@ PARALLEL\_IO\_MODE | netcdf4, hdf5, pnetcdf | mpiio and mpiposix are deprecated
 -----------------------|--------|-----------------------------------------------------------
   FLUSH\_INTERVAL       | int   | Minimum time interval between flushing heartbeat data to disk.  Default is 10 seconds
   FLUSH\_INTERVAL       | int   | For non-heartbeat, the number of output steps between flushing data to disk; if 0, then no flush
-  TIME\_STAMP\_FORMAT    | [%H:%M:%S] | Format used to format time stamp.  See strftime man page
+  TIME\_STAMP\_FORMAT    | \[%H:%M:%S] | Format used to format time stamp.  See strftime man page
   SHOW\_TIME\_STAMP      | on/off | Should the output lines be preceded by the timestamp
-  PRECISION            | 0..16 [5] | Precision used for floating point output.
+  PRECISION            | 0..16 \[5] | Precision used for floating point output.
   FIELD\_WIDTH          | 0.. |  Width of an output field. If 0, then use natural width.
-  SHOW\_LABELS          | on/[off]  | Should each field be preceded by its name (ke=1.3e9, ie=2.0e9)
-  SHOW\_LEGEND          | [on]/off  | Should a legend be printed at the beginning of the output showing the field names for each column of data.
-  SHOW\_TIME\_FIELD      | on/[off]  | Should the current analysis time be output as the first field.
-
+  SHOW\_LABELS          | on/\[off]  | Should each field be preceded by its name (ke=1.3e9, ie=2.0e9)
+  SHOW\_LEGEND          | \[on]/off  | Should a legend be printed at the beginning of the output showing the field names for each column of data.
+  SHOW\_TIME\_FIELD      | on/\[off]  | Should the current analysis time be output as the first field.
 
 ## Experimental
 
  Property              | Value  | Description
 -----------------------|--------|-----------------------------------------------------------
-MEMORY\_READ        | on/[off]   | experimental
-MEMORY\_WRITE       | on/[off]   | experimental
-ENABLE\_FILE\_GROUPS | on/[off]   | experimental
+MEMORY\_READ        | on/\[off]   | experimental
+MEMORY\_WRITE       | on/\[off]   | experimental
+ENABLE\_FILE\_GROUPS | on/\[off]   | experimental
 
 ## Debugging / Profiling
 
   Property | Value    | Description
  ----------|----------|------------
- LOGGING   | on/[off] | enable/disable logging of field input/output
- DECOMP\_SHOW\_PROGRESS | on/[off] | show memory and elapsed time during autodecomp.
- DECOMP\_SHOW\_HWM      | on/[off] | show high-water memory during autodecomp
- IOSS\_TIME\_FILE\_OPEN\_CLOSE | on/[off] | show elapsed time during parallel-io file open/close/create
+ LOGGING   | on/\[off] | enable/disable logging of field input/output
+ DECOMP\_SHOW\_PROGRESS | on/\[off] | show memory and elapsed time during autodecomp.
+ DECOMP\_SHOW\_HWM      | on/\[off] | show high-water memory during autodecomp
+ IOSS\_TIME\_FILE\_OPEN\_CLOSE | on/\[off] | show elapsed time during parallel-io file open/close/create
  CHECK\_PARALLEL\_CONSISTENCY | ignored | check Ioss::GroupingEntity parallel consistency

packages/seacas/MAPVAR.md

@@ -0,0 +1,44 @@
+After a look and run through the debugger; I'm not sure how the nodal
+variable interpolation *ever* worked in the past.  It looks totally
+wrong and I can understand why it isn’t working now; I just can't
+understand how it ever worked correctly for multi-element block
+models.
+
+As I see it, it does the following for the interpolation:
+
+ * For all blocks
+
+   * For all time steps
+
+     * For all nodal variables
+       * Iterate all nodes in this block; map from A->B
+       * Write values for *all* B nodes at this step for this variable
+
+   * For all time steps
+
+     * For all element variables
+       * Iterate all elements in this block; map from A->B
+       * Write values for all elements in this block at this step for this variable
+
+This works for element variables since the exodus API can output
+elements a block and variable at a time.  For nodes, it doesn't work
+and you will end up with the values at the last step for all
+nodes/variables except for nodes which are only in the last element
+block which seems to be what you are seeing.
+
+Fixing this would be a major undertaking and I'm not sure it would get
+prioritized (although you are welcome to try).
+
+This *should* work OK if you only do a single timestep or if you only
+have a single element block.  With a single timestep and multiple
+element blocks, there is an issue of what happens if the node is
+shared between multiple element blocks -- it will only get the
+interpolated value from the last block.
+
+Now, what to do...
+ * I think that the Percept code can do some mapping from mesh to mesh...
+ * Klugy, but can do a timestep at a time and then rejoin all timesteps using `conjoin`
+ * Klugy, but can subset down to one block / mesh and then run mapvar on each submesh and then join using `ejoin`
+
+Sorry for the bearer of bad news, but hopefully there is a path to get
+done what you need...

packages/seacas/NetCDF-Mapping.md

@@ -29,8 +29,7 @@ used in an Exodus file.
 * There are about 10 standard dimensions in every file.
 * plus one for each set plus one if any attributes
 * plus two for each block plus one if any attributes
-* plus one for each transient variable on an entity (node, node set, element
-block, element set, ...)
+* plus one for each transient variable on an entity (node, node set, element block, element set, ...)
 
 ## Variables: (NC_MAX_VARS)
 * There are a few standard dimensions
@@ -41,19 +40,25 @@ block, element set, ...)
   * #ndim coordinates (1,2,3)
 
 * Each block adds 1 + 2*#attr_on_block + #var_on_block
+
 * Each set adds 2 + 2*#attr_on_set + #var_on_set
+
 * Each sideset add 3 + 2*#attr_on_sset + #var_on_sset
+
 * Each map adds 1
 
 ## Example
 If we have an exodus file with:
  * Nodes
+
  * 5 Element blocks
    * 4 transient variables per element block
    * 2 attributes per element block
+
  * 4 Node Sets
    * Distribution Factors defined on each set
    * 3 transient variables
+
  * 3 Side Sets
    * Distribution Factors defined on each set
    * 2 transient variables