I think, the term “flatten” is mostly applicable to raster media, when PDF is a vector format. I have no idea what would mean to flatten, for example, two line drawling commands line(0,0,100,100) and line(0,100,100,0)…

PD4ML allows to convert source HTML to “flatten’ raster images, but in print scenarios the approach would make not much sense, because of print quality losses and bulky files.

I would suspect the slow print is caused by an exceeding of printer memory with big PDF print jobs. How big an average PDF file in the problem cases?