Disregard previous post. Identified and fixed. (Parameter is controlled by an uncalibrated slider in Cura window.) Also did not solve my Gcode mismatch probelm, so we will move on to that.
When I compare old and new Gcode files I find:
All X, Y, Z positions match perfectly.
All feedrates (F term) match perfectly.
All S terms match perfectly.
All block sizes, block identifiers, etc, match perfectly.
The line count matches (that is, a line of Gcode in the new file is in exactly the same place vertically from the start of the file as in the old (baseline) file.)
The filament feed is low by about 9% in the new file compared to the old file. The E term is cumulative, but it shows up in the very first print line, then degrades by the same proportion as the part generation proceeds. I copied the CuraEngline parts of the Repetier log out form old and new machine. The logs match perfectly except at the end where the total filament used is reported. The new machine generation report is about 9% lower than the old report. That is consistent with what I see in the incremental feed values.
I believe that this feedrate is determined within Cura but depends on some parameter sent over from Repetier when Cura is launched. Cura uses an .ini file to control its configuration, but I copied over the old .ini into the appropriate Cura directory.
Any thoughts on this? Right now, the output will underfeed filament into the part by 9%.
Feel free to correct my understanding of the process.