I start Julia executable (called by a script) that execute code and exit.
Benchmark is useful to better study some performance oddities resulting from executions in Julia 0.5 (last stable version when code was created) and Julia 0.6 (actual last stable version) - see another my post
Leonardo