hello, it appears that windows 7 build 6509 from quick testing does not seem to suffer from this issue although 6519 has been claimed to have fixed this bug first.
i tried many reboots (i think i did it 20-30 times or more) and not one time has it failed unlike the known build before it (build 6498) where it once failed to start 1 service or it displayed a black screen upon boot or the setup just didnt launch.
maybe this information can help to narrow down where the drift bug fix first got implemented before other changes have occurred that may have muddied the waters in the code.
maybe its worthy to those that are still working on a solution. (win32ss and maybe some others)
please note that i have a haswell machine (i5 4460).

Edited November 16, 20241 yr by ward201185

8 hours ago, ward201185 said:

please note that i have a haswell machine (i5 4460)

Compare it to the behavior of standard Windows Vista

There is a chance you got very lucky in the silicon lottery and your particular CPU doesn't suffer from the drift bug.

I already tried windows vista and no, im definitely not lucky at all in the silicon lottery otherwise i wouldnt have brought this up.
Im no stranger to the problems on win vista for my cpu generation and i still remember all the side effects i had when i tried vista x64 but i didnt notice anything different with build 6498 for that matter. 
It was just odd that build 6509 had survived ALL the reboots i made and i have yet to experience any of the issues i had on win vista x64, it worked perfectly just like my vista x86 install which definitely doesnt have this problem at all. 
Seriously, try it and see for yourself, you'd prob see a difference between build 6509 and 6498 so make a usb stick of 6509 and try to swap boot.wim periodically after some amount of restarts with a regular vista boot.wim and/or build 6498 boot.wim and you'll see what im talking about. 

Edited November 17, 20241 yr by ward201185