The 2 computers are off about 7% between the two. One is higher than the other. The 7% difference is calculated from an average of many of the same rides. Each bike's ride individually is very close and repetitive (if I take the upper and lower distances of each bike I get about a 2% difference ride to ride span).
What I don't like about the "measuring the circumference of the wheel" method is this does not take in to account the compression of the tire when under load.
That is why I am trying to get an accurate course distance measurement. Then I can tune the calibration factor on the computer per bike so that the reading is repetitive and correlates to the measured course.