parallel_sfs is a wrapper around
(Section 16.61) meant to divide the input DEM into tiles with overlap,
sfs on each tile as multiple processes, potentially on
multiple machines, and then merge the results into a single output
DEM. It has the same options as
sfs, and a few additional ones, as
Examples for how to invoke it are in the SfS usage chapter.
parallel_sfs -i <input DEM> -n <max iterations> -o <output prefix> \ <images> [other options]
Command-line options for
- --tile-size <integer (default: 300)>
Size of approximately square tiles to break up processing into (not counting the padding).
- --padding <integer (default: 50)>
How much to expand a tile in each direction. This helps with reducing artifacts in the final mosaicked SfS output.
- --processes <integer>
Number of processes to use on each node (the default is for the program to choose).
- --num-processes <integer>
--processes. Used for backwards compatibility.
- --nodes-list <filename>
A file containing the list of computing nodes, one per line. If not provided, run on the local machine. See also Section 8.16.
- --threads <integer (default: 8)>
How many threads each process should use. This will be changed to 1 for ISIS cameras when
--use-approx-camera-modelsis not set (Section 16.61), as ISIS is single-threaded. Not all parts of the computation benefit from parallelization.
Resume a partially done run. Only process the tiles for which the desired per-tile output files are missing or invalid (as checked by
Suppress output of sub-calls.
- -v, --version
Display the version of software.
- -h, --help
Display the help message.