vbmm2

vbmm2 carries out various matrix operations at the command line, on VoxBo matrix files only. It also allows you to do most of these operations in parts. It's mostly for internal use in VoxBo GLM job sequences. Here's the help text:

VoxBo vbmm2 (v1.8.5/Jun 14 2011)
summary:
  vbmm does various bits of matrix arithmetic with matrix files.
  below, in and out refer to the input and output matrices.  c1 and c2
  refer to the start and end columns to be produced (for parallelization).
usage:
  vbmm2 -xyt <in1> <in2> <out> <c1> <c2 >     do part of XYt
  vbmm2 -xy <in1> <in2> <out> <c1> <c2>       do part of XY
  vbmm2 -imxy <in1> <in2> <out>               I-XY in core
  vbmm2 -f3 <v> <kg> <out>                    V*KG*invert(KTtKG)
  vbmm2 -xyz <in1> <in2> <in3> <out>          XYZ in core
  vbmm2 -assemblecols <out>                   assemble out from available parts
  vbmm2 -assemblecols <in1> <in2>... <out>    assemble all in files into out
  vbmm2 -assemblerows <out>                   assemble out from available parts
  vbmm2 -assemblerows <in1> <in2>... <out>    assemble all in files into out
  vbmm2 -add <in1> <in2> <out>                add two matrices
  vbmm2 -subtract <in1> <in2> <out>           subtract in2 from in1
  vbmm2 -invert <in> <out>                    invert in
  vbmm2 -pinv <in> <out>                      pseudo-inverse, in core
  vbmm2 -pca <in> <out>                       calculte pca
  vbmm2 -ident <name> <size>                  create an identity matrix
  vbmm2 -zeros <name> <cols> [rows]           create a zero matrix
  vbmm2 -random <name> <cols> [rows]          create a matrix of random numbers
  vbmm2 -print <name>                         display a matrix
  vbmm2 -printsub <name> <r1> <r2> <c1> <c2>  display part of a matrix
  vbmm2 -compare <mat1> <mat2>                compare two matrices
  vbmm2 -reshape <in> <out> <n> <rows/cols>   reshape an input vector into a matrix
  vbmm2 -applyfilter <in> <filter> <out>      apply "ExoFilt" filter to vector
  vbmm2 -deriv <in> <out>                     calculate derivative of a vector
notes:
  -assemblerows and -assemblecols, given a single argument, will try
  to find the pieces for that file, and will delete those pieces when
  done.  Given more than one argument, all but the last argument will be
  taken as input filenames, and the last will be used for the output.

  -reshape allows you to fix the number of rows or columsn.  if you fix
  the rows, the vector is laid out column-by-column.  if you fix the cols,
  the vector is paid out row-by-row.