Figure 8 on Page 6 of the actual paper shows what they're measuring. They're comparing filter materials by Salt rejection % vs Water permeability measured in L/cm2/day/MPa. That unit incorporates all the energy-efficeny goodness you want in a filter without looking at what pump technology is actually used to provide the energy input. It says that more filtered water (L) per square centimeter of filter (/cm2) per day (/day) per MegaPascal of pressure (/MPa, the energy input) is more good. Assuming any particular pump technology would give you a number for MPa/MJ that you could apply, but it doesn't help you understadn the performance of the filter itself. The figure for improvement vs existing technology they actually give is 2-3 orders of magnitude (100-1000x) so TFS is taking the optimistic side.
The bottom line is that this has a huge potential but is still a ways from practical application.