Reassigning Tracks vs Processing

I have always been partial to offline processing media when I was editing. Especially, back when I started editing. The advantage of A-B comparison and highly tuned parameters were all I could see. I would spend hours after an edit processing individual clips or scenes. Sometimes to my detriment.
As someone wise once told me you need to see the forrest through the trees; in editing as in life.
Enter the question: which workflow is better when working with denoisers? The answer is disputed in friendly both are valid workflows it is a matter of the application of reassigning tracks vs. processing.

Back then, online processing required the suite you were cutting in to be particularly beefy, with CPU clock speed being a major factor. That isn’t so any more, with the newest algorithms and high speed i7 processors. It was time for me to reevaluate my notions of online processing.
I would spend hours on a returning pass of dialogue, processing clip by clip sometimes just to maintain the level of noise floor. It was painstaking work requiring processing and endlessly comparing the resultant files to the originals. The start/stop nature of the work is tedious. Especially the indecision of inexperience.
The process of setting a track to handle the noisy media, allocating all noisy media to that track and hoping the machine wouldn’t choke on it seemed too cludgey. I would take pride in a dialogue edit that played like a mix. I had a luxury then that on most of the shows I have cut or mixed since, hasn’t been available to me: ample time. Now, I see how that even allocating can be time consuming.
When handling someone else’s tracks, that hadn’t taken denoising into account reassigning can be a god send, saving tons of time.
At this time I believe all the major DAWs for film post will allow for automatable bussing reassignment. If you are stuck in the old workflow try this instead.
The process of reassigning busses in pro tools is version specific. I believe that HD versions of pro tools allow you to reassign busses easily. In the non-HD version you must make a null auxiliary track to work around this. Prefader sends work but are a bit more cumbersome in operation. You must make sure that mutes (in your preferences) are not pre-fader or that your fader is down. Every time you automate the bussing of a section you must be sure not to double bus the signal. With bus reassignments you simply loop the section as normal and select the alternate bus.
That said if I am given ample time to edit I still go for processing as my edit suite is still a Q9550 series Intel quad core processor. In a complicated template the VST performance meter has hardly any headroom and to throw another algorithm in the mix introduces clicks and pops and stalling on playback.
Conclusion:
The truth is no matter which approach you take it is always important to compare the original audio against. your processes. I find in the online workflow it is easier to leave the processor turned up a bit high or engaged when it need not be.