LAMMPS-style files are often quite large and can take up a lot of space. To recursively compress all dump files in parallel, run:
find . -name "*.dump" -type f -exec pigz {} +This will find all files with a .dump prefix in the current directory (including those in subdirectories) and compress it using pigz, which is, crudely, just a resource-aware gzip, i.e. it will run compression algorithms in parallel.
Running into an error about some illegal character ^M? This is likely an artifact from editing the file on Windows. There's a handy command for converting DOS (Windows's format) to Unix:
dos2unix file.txt