three thoughts...
- i make a folder with jpeg2000 files & 5.1 audio and then wrap them into a DCP with DCP-0-Matic, advantage is when a small change is requested (ie: a logo added to the head or a .com added to the tail) i only need to render the frames changed and re-wrap - the re-wrap takes about 7-12 min on my machine, it's a file copy and write a new XML, so dependant on disk speed mainly
- i have dual 2697v2's and and typicaly run a 4k Jpeg2000 seq in about realtime, 2k seq runs around 40-60fps with all threads flat out, so the pain of 11 hours of rendering and then another 11 hours of re-rendering is not something i have to deal with
- v16's flexiable timeline raster should help with your workflow, i typicaly have credits at both 4k and 2k on seperate video tracks, and bypass what ever i'm not exporting, but avoiding scaleing credits means seperate timelines anyway, in v15 it means copying the project and resetting the raster to 2k... then lineing them both up to go at the same time, no huge advantage to v16, but it does save a step or two and makes keeping track of changes easier