After the May update the difference is smaller, but there is still a difference. Both INF and POJB have differences between BK and T7 versions. Since the reports are being run for auditors there is no wiggle room. They demand to know why there is a difference, which one is correct, and how we are going to test the numbers to prove them. They will be disappointed with my answers, no matter what I come back with, so I'm not too worried about that. The numbers are garbage anyway because I have not run SM-J-C since last June.
Evo is prone to non-transactional changes to UOH, so relying on as-of dates if a bit silly beyond a few days.