OK now I've eliminated some drifting in RCT <- yes RCT. I've monitored both gen1 and gen2 vm for something like 12 hours. no drift in 12 hours of VM idling.
this doesn't solve the thing: issued a reboot. got a shutdown! but at least without timing integration services now clocks seems logic in the VMs. I think even @JaredBusch got issues with timing and constant time reassignments...
I've also rebooted the server from the closet, nothing strange in the reboot sequence: all ok with the hypervisor. I've a new win update. let's apply this. also let's check the "latest" HPE service pack from october 2016, maybe some driver is apparently ok, but requires update...
bah. I'm going to think it's me. I'm cursed w/ hpe stuff.
A last thing remains: enter bios disable power savings and CPU c-states and let's see... then I'll STONITH my self
f#@*$%!