Might also wanna note the workaround, which is setting the car_rate to 115,200.

If Mark's explanation is correct (i.e. that stutter is caused by waiting for kernel to write debug message on slow serial line, as opposed to stutter and message having common cause), I am affraid this will alleviate the problem only partially, since I am quite reliably getting a bit of stutter at home as well. (Actually, this can be a test of Mark's explanation: if writing on serial is not the cause, then changigng the rate won't help at all.)