Caching Data For Searches

Searching large numbers of files for similarity can take significant cpu time. Part of this is the time to compute the similarity matrix for each file.

If the Cache Thumbnails option is selected in Preferences/General, the similarity matrix and the checksum will also be cached. This will reduce the time needed for future searches.
If you frequently search on similarity and your images are in a tree arrangement under a single point, initiating a one-time search on similarity from the top of the tree will generate the similarity data for all images.

Similarity data are stored in a folder hierachy that mirrors the location of the source images. The data have the same name as the original appended by the file extension .sim.

The root of the hierachy is:

or, if $XDG_CACHE_HOME is not defined: