Easier test requiring no extra equipment:
Run both machines in record mode simultaneously but seperately for a signifcantly
extended time (like over night for 12 hrs), with a clap at the start and end of the test recording. File splits shouldn't matter.
Do the same again with them linked/slaved together.
Align the files and check for drift. Unless the clocks are actually locked, there is no way you would see zero drift down at the sample level over the course of a 12hr recording even if the clocks were so closely matched that the difference isn't detectible with a typical ~2 hour music recording.
[edit- a beter solution than linking R-44s is an R-88.
]