LArRTMExtractor, LArOFCAlg: Allow to choose the number of threads when executing with tbb
The algorithms LArRTMExtractor and LArOFCAlg do all their CPU-intensive work in stop(). This work can be spread out on many cores using tbb::parallel_for (this has nothing to do with athenaMT in general). With this MR, the number of threads becomes configurable.