[ros-dev] RE: [ion] 19742: -ImplementRtlPrefectMemoryNonTemporal.Patch by PatrickBaggett<baggett.patrick@gmail.com> and myself.

[Ge van Geldorp]

> From: Alex Ionescu
>
> This is a very good point, but I think there are much bigger 
> problems then this to get NT loading on a ROM.

So that's an argument to introduce code which brings us further from the
goal?

> This routine is pretty new and that's why it's not called, 
> but from what I've read on SSE, prefetching is supposed to be 
> an ultra-time critical operation

Then inside the kernel it should be inlined, saves a CALL and RET. For use
by drivers it should be declared as __attribute__((dllimport)) in the
headers, saves a JMP. That is, if I buy the "ultra-time critical operation"
which I still do not, without any timing evidence.

> When I studied the routine under Windows and I saw that a 
> team of over 2 500 hardware and software experts with access 
> to gigantic user bases, profiling equipment and connections 
> with developers and Intel made the same decision, it was 
> gratifying.

How can we claim that ReactOS is an independent implementation when we start
to use "well, I reverse engineered Windows and Microsoft does it the same
way" as a implementation decision rationale? For that matter, why spent time
on ReactOS at all? Obviously, Microsoft with its huge teams knows better
anyway, so we should just stick to Windows.

GvG