A couple interesting points ChatGPT in research mode found regarding the potential benefits of parallel:
Here's the full result, but note that it seems to not be grok'ing the idea that in parallel the proposal was run the flow at half speed, so it's giving incorrect info about the dwell time...