your boygroup_without_LFO example is typically the one to go for. as in the other example independent lfos run on the clients and they are not guaranteed to run in sync forever after you started/reset them. even starting/resetting them via a bridge (=connection to a non boygrouped node) to a bang is not guaranteed to start them at the exact same time.
therefore sending one master-time to all clients and doing local smoothing is the way to do it.