Laminar to molecular flow in a vacuum piping system, which will work better?

Let's see if I get an answer faster online than I do from my department engineers. In a given length of pipe with minimal bends, the pressure is dropped from 760 Torr (7.6 X 10^2) ((6 millionths of an inch Mean Free Path)) to 7.6 X 10^-9 (30 miles MFP) In transitioning from laminar to molecular flow, we have a stalling point where our mechanical pumping method is not efficient, resulting in huge amounts of time ($$) being wasted. Which will work better, heating the pipe(ing) to half melting point (Stainless 316L) to "cook out" random molecules of moisture thus dropping the pressure OR cooling the pipe with cryogenic traps (-196C) to "freeze" the same molecules to the inside of the pipe and reduce the pressure that way. $64,000 question. (More actually) You are up against 8 real life people (6 men, 2 women) that have been working this since Monday 9/22/08. Not going to kill you with formulas for adsorption (yes with a D) Just curious to see what the consensus is from a large sample of engineering students, at least the ones that respond, then compare with the report when it comes in.