So you can calculate, for a given thrust/energy ratio, what velocity you have to exceed to get more energy out than you put in. For relativistic effects to make

So you can calculate, for a given thrust/energy ratio, what velocity you have to exceed to get more energy out than you put in. For relativistic effects to make enough difference to matter, that velocity would have to be pretty close to c. That translates to an extremely low thrust. (I've been told that for a photon rocket, the threshold velocity is c, implying that a reactionless drive won't violate conservation of energy as long as it's no better than a photon rocket. I haven't done that math myself though.)