User Tools

Site Tools


jvs:cigal:manual:chapter5:dmformat

DM archive file format

CIGAL uses a data management (DM) archive file format to store many different types of data together in a single self-documenting file. Data records can be binary (8-bit, 16-bit, 32-bit integer or 32-bit floating point in big or little-endian format) or text. Binary records are stored as 1D, 2D, or 3D data arrays. All data records in the file are simply catenated one after another (a few padding bytes may be added between records to ensure that every record begins on a 4-byte boundary). The file is described by a text header. There is usually a fixed format header at the beginning of the data file itself, or the header can be contained in a separate file.

The fixed format header has the following form:

DMDATA     8192 abcd
#     Type Wsiz      Address           NX     NY     NZ     NT
Record    1   0         8192        12915      1      1      1
  Text: Paradigm prog: showplay
Record    2   0        21112         4446      1      1      1
  Text: paradigm parameters
Record    3  32        25560            2   5742      1      1
  Text: run-time log
Record    6   0        71496            3      1      1      1
  Text: comment
(more Record/Text pairs for each record in the archive)

The record TYPE parameter identifies the type of each record. An application (e.g. CIGAL, fScan) uses unique TYPE values for each type of record, regardless of where the record is found in the archive.

The WSIZ parameter specifies data word size format. Values can be:

  0 - ASCII text record
  8 - 8-bit binary integers
 16 - 16-bit binary integers
 32 - 32-bit binary integers
544 - 32-bit binary floating point
jvs/cigal/manual/chapter5/dmformat.txt · Last modified: 2023/02/23 18:43 (external edit)