youtube[.]com/watch?v=izqEZmjTfuM
neowin[.]net/news/report-your-windows-11-pc-can-silently-cripple-itself-and-you-wont-even-know-it/
Lower Performance from a *Bad* copy of Windows, and inconsistent results from Win11 KB5041587 & Win10 KB5041582
As I guess I found out upgrading my Windows To Go drive to Win11 24H2, where the poor performance of the upgrade prompted a fresh install, there is apparently such a thing as a good Windows install and a bad one [24H2 requires reinstalling Windows]. And this can effect CPU performance, particularly with AMD Ryzen CPUs, especially with more demanding tasks like gaming. The 2 KBs referenced add Microsoft's patch for Ryzen CPUs, which can have a "meh" or "Oh WOW!" effect, depending on the game, and if you had/have a bad Windows install. AFAIK there are no public details on what the patch actually does, and no details on why it boosts one game more than another, but logically you should see the same variable gains in non-game software too. I felt that the fact that there are good & bad Windows installs warranted a new post & thread. While the patch does help *bad* copies of Windows, there's not enough info &/or data to tell if it really fixes them, or if there are more underlying problems. Likewise, while the problem seems more pronounced, or maybe more common with Win11 23H2, does it occur with other Windows versions as well? And finally, how common are bad Windows installs?
In the video linked above, a reviewer stumbled across the problem of a bad Windows install while running benchmarks, trying to make sense of the Big variance between different reviews of the new Ryzen CPUs. For each set of benches he installed Windows fresh on a new NVMe SSD. Since he just swapped out the SSDs, he could go back and retest when his results were confusing. Most of the video documents the benchmarks with Win10, patched & unpatched, Win11 good & bad installs, again patched and unpatched, and Win11 24H2. Through it all the test PC remained unchanged.
Win10 & 11 both suck compared to Rock Steady Win7 [and *maybe* Win8 too, but I never used it enough to say one way or the other]. Back in 2015 I complained about Win10's lack of consistency -- with Win7 EVERY copy of Windows performed the same, and what worked, what you saw with one copy was exactly what you saw from every other copy [unless something was broken]. With Win10, from day one, what you experienced with one copy may or may not be the same with the next copy of Win10 you used. This inconsistency has continued with Win11, and now it seems is more serious than I had thought.