Not positive. Not sure what the difference is either. I would think a timedemo is a timedemo. As long as the same parameters are maintained, then benchmarking is consistent. Technically speaking, any timedemo would work really, as long as you run the same one on different settings and/or hardware to gauge the scores and differences in performance.
Any particular reason it has to be a specific timedemo? Are you trying to compare your results to Tom's? If so, I'd caution against it......comparing like that requires EXACTLY similar setups in order to get true, unbiased results. Very tricky.