As I generally print my better images in two paper sizes - 13x19 and 17x22-25 - I have followed Schewe's past advice of upsampling first in ACR (don't use LR), taking it to the max. of ~28mp/16-bit PPrgb for my 16mp files, and de-convolution capture sharpening. I output the 300ppi TIF's to PS6 perfectly sized for my ~15.5x20.5 uncropped images, finishing with customized Smart Sharpen, often first with small radius, then wide radius-low level for local contrast enhancement.
Lately I've been in the habit of simply resizing without downsampling for my 13x19 (12x16) output, ending up printing these at around 380ppi. Heresy, I know, but I'll be hanged if I can see any IQ difference from my iPF5100 output with the suggested 300ppi downsampling. Printing in highest accuracy 600dpi mode, through the excellent 16-bit Canon plug-in. I've compared at my very myopic 6" focal point without glasses, but not loupe, and haven't compared upsampling to 600ppi for the 13x19's.
Anyone else with experience using this "one file fits all sizes" approach, particularly with newer printers?