=============================== ChangeLog ERF-HDF5 2.1 -> 2.2 =============================== - new block type "properties" key=500 added The new "properties" block is a general purpose block to store integer, floating point and character data for any kind of entities. Currently, the VPS solvers use the "properties" block to store data for result picking. =============================== ChangeLog ERF-HDF5 2.0 -> 2.1 =============================== - new block type "identifiers" key=250 added The new "identifiers" block is a clone of the "attributes" block. It just comes with a different key and block type string. The purpose of this block is to store user identifiers and module ID's to support modular solver I/O. Note: Modular I/O is a new VPS solver feature to enable the user to easily exchange or include model components without taking care of the uniqueness of the user entity IDs (nodes, elements, parts etc.). The solver internally uses sequential numbers to avoid clashing ID's between the main and the submodels. These sequential numbers are written as entity ID's in the ERF result blocks and can be retranslated into pairs of user and module ID's by using the information stored in the new "identifiers" blocks. =============================== ChangeLog ERF-HDF5 1.2 -> 2.0 =============================== General: -------- - The three main ERF-Block groups have been merged into one single group, called "erfblock": "header", "data", "index" -> "erfblock" - A new datatype for integer identifiers have been added. There are five types of data now: INT (no change) - for integer values, flags, numbers FLOAT (no change) - for real values, results CHAR[n] (no change) - for strings LONG (no change) - for array pointers, counts, offsets INTID (new) - for integer entity identifiers The type INTID has been introduced to have the possibility to declare the entity IDs either as 4-byte (standard) or as 8-byte integer for future extensions. - Some data were converted from HDF5-datasets into HDF5-attributes. In the ERF specs, it is indicated in the column "type": * The prefix "D:" stands for an HDF5-dataset and * the prefix "A:" stands for an HDF5-attribute. Examples: A:INT, D:FLOAT The new parameter "indexic" of the simulation state index (all indexable ERF Blocks): -------------------------------------------------------------------------------------- - The parameter "indexic" was added to flag blocks containing incremental data. This new feature will be used for large output data with very fine discretisation and many states. The first application using this new feature is the ProCAST/CAFE module. CAFE visualizes misorientations of grains in very fine FD-grids. Note: The flag indexic is only available for single-state - entity results, - element results, - attribute and - active-flag blocks. Variable Definition: -------------------------- - The "variables" block has been splitted into a "variablegroup" and a "variable" declaration block - The "variablegroup" block contains one or multiple variables, each identified by a varkey string. The string etypvar now represents a GROUP of variables. Note: varkey[nvar] is an array of nvar strings. The string varkey is a variable key that is unique over all variable group definitions, - The "variable" block contains a variable definition. Each variable is identified by the variable key that is stored in the string varkey. The string varnam contains a name to label a variable (e.g. in a menu or on screen). The variable map definition vmap replaces the old comp arrays and is used to specify the symmetry conditions of variables of rank > 0 (tensor or matrix). The new subscript array (character array subscr) replaces the cname array. A subscript can be attached to any variable coordinate. Subscripts are supposed to be used as suffixes of the variable names. The new variable class string (character array varclass) is used to classify the variables, e.g. for compression filter settings. Examples of classes are: "CLASS_LENGTH", "CLASS_STRAIN", "CLASS_STRESS" etc. Connectivities Block: --------------------- - In the "connectivities" block the new parameter "etyp" was added (in addition to the parameter "etypelem"). The parameter "etyp" contains a generic entity type to address a category or family of different entity types. For instance, etyp="ELEMENT1D" would stand for any type of a finite beam or bar element. The advantage is that element results of different element types can be stored together in the same "entityresults" block using the identifier etyp="ELEMENT1D". - In the "connectivities" block the variable definition etypvar was added to define the names and the units of the optional fparam values. Hence the unit definitions have been removed from the "connectivities" block, because they are assigned in the variable definition block. Parts Block: ------------ - In the "parts" block the array pvtyp was removed. The arrays "mid", "mtyp" and "pcol" are now optional data. The new parameter "off" was added for convenience. It can be interpreted bitwise to find out which of the arrays "mid", "mtyp" and "pcol" does NOT exist. For instance, off=0 means all arrays are stored. - The length of the title string (array "title") is defined by the parameter "length". The parameter "length" can be greater or equal to zero. length = 0 means there are no titles stored in the block. Attributes Block: ----------------- - The length of the title string (array "title") is defined by the parameter "length" (similiar to the "parts" block). The parameter "length" can be greater or euqal to zero. length = 0 means there are no titles stored in the block. Text Block: ----------- - In the "text" block the optional state index was added. Indices Block: -------------- - In the "indices" block the new variable definition etypvar was added to declare the meaning of the index values. Hence the unit definitions were removed from the "indices" block.