So if I had my guess windows 11 will fuck up drivers that worked fine in 10 like every successive major release has done

So here’s a good question
What the hell do they change every time that causes this ?

Ofcourse they will

@EpicofGilgamesh yes but what do they do that that happens ?

I would bet that the core is actually exactly the same with just minor tweaks, except the new directx drivers (weren't they available for win10 as well?)

I heard theyre targeting only newer cpus, which would allow them to presumably simplify their kernel a bit compared to the wider support seen in 10.

Would it make windows 11 faster? Depends on if they werent already bundling redundant binaries to account for optimizations available in newer cpus but not older ones.

Would it make windows 11 smaller? Depends on the above and whether savings on one side of the OS are being used to justify filling up that space some other way.

Does it require a 10->11 change?
Yeah unless windows wants to brick a large number of devices post update.

Will it make their hardware partners happy? Definitely.

So here's a question. how many differing features do differing architectures offer OBJECTIVELY ?

Like processor set shouldn't change much these days i would think.

its mostly amd and intel on the market.

you're either receiving data from a device in some port of memory, or handling an interrupt from a device, or sending an interrupt or input to a device, or executing code in ram, or directly from a disk device correct ?

god i remember being cheerful when their seemed to be a point to this crap and their bullshit wasnt so obvious.

oops i didnt say that :P

@MadMadMadMrMim I guess, mostly power draw. ARM is more power efficient, but has a much simpler instruction set. The same goes to RISK-V which is a more basic set of instructions than x86, but because of that simplicity should be a much lesser power hog, while being as effective.

@MadMadMadMrMim btw, you have forgotten ARM. A ton of stuff is actually powered by ARM based processors.

So i found an article on this but it looks like its actually about official support rather than a hard restriction so i was wrong

https://theverge.com/2021/6/...

Looking elsewhere it looks like the breaking hard requirement changes are more around:
- requiring 64 bit processors
- hardware level security via TPM
- requiring more ram (not good)

Back to the question around CPU features, its more about instruction sets.

When a new intel processor is released it could hypothetically have a new supported instruction that when used speeds up particular computations. But to make use of that instruction your binaries themselves need to reference it, which means those same binaries cant be executed by processors lacking that instruction.

Thats my understanding of it anyway

They change the version string so a bunch of asserts randomly fail.

@homo-lorens asserts in compiled drivers ? is that what you mean ?

is that a joke ? or are you serious ? because like I've seen system crashes from old drivers.

@MadMadMadMrMim It was a joke, but runtime environment assertions in dynamically loaded drivers that don't crash the system if they fail aren't unheard of.