Run the following commands in the source directory, be sure to set the MPICC variable to your c++ mpi wrapper:

autoreconf -iv ./configure MPICC=mpiCC make

The default binary is a simple program that will reproject rasters in parallel using mode statistical resampling. The program takes the following parameters:

src/prasterblaster [-n <partition count>] -p <output projection srs> [-f <fill value>] [-r resampler] <input raster path> <output raster path>

• -p, –projection=

  The projection is specified using a proj.4 parameter string. For example:

  “-p “+proj=latlong”” specifies and unprojected dataset.

  “-p “+proj=moll”” specifies the Mollweide projection.

• -n, –partition-count=

  This parameter specifies the number of partitions to divide the rasters into.

  For example: “-n 500” specifies 500 partitions

• -f, –fill-value=

  This parameter specifies a value to assign to pixels outside of the projected area.

  For example: –fill-value=100

  Specifies the fill value to be 100. How the value is interpreted depends on the type of the raster pixels. I.E. you specify a float and the raster value is integral the fill value will be truncated.

• -r, –resampler=

  This parameter specifies a resampler to use to calculate output values.

  Possible values: Min, Max, Mean, Mode, Bilinear,

• <input raster path>, <output raster path>

  These are file paths that specify the locations of input and output rasters. The input must be in geotiff. The output will be written in geotiff.

  For example:

  tests/testdata/input.tif

• -t, –temporary-path

  This parameter specifies a path that per-process temporary files will be store in. This path should not be network shared.

src/prasterblaster -f 254 -r Min -n 100 -p “+proj=moll” -t tmp \ tests/testdata/glc_geographic_30sec.tif \ tests/testdata/glc_mollweide_30sec.tif

This will produce a raster in the Mollweide projection.