I was trying to use a CPU detection routine that relied on the
undocumented UMOV instruction to differentiate between AMD 386sx and
Intel 386sx chips, but testing that routine has found it flawed (ie.
everything is coming up AMD even on parts marked genuine Intel).  The
routine assumes that UMOV is present in AMD chips but not in Intel.
The actual routine is reproduced here for the curious:
http://www.vintage-computer.com/vcforum/showthread.php?28026-Looking-for-volunteers-to-help-test-a-new-benchmark&p=205608#post205608

After several tests by several people, it appears this assumption is
*incorrect* and may actually be *reversed* (meaning, Intel chips DO
have UMOV while AMD chips do NOT).

Can anyone clarify this?  Or, is there a better way to determine if a
386 is an AMD or Intel part?

On Dec 30, 9:50=A0am, Jim Leonard <mobyga...@nospicedham.gmail.com>
wrote:
> Can anyone clarify this? =A0Or, is there a better way to determine if a
> 386 is an AMD or Intel part?

I think I found my own answer at pages that are no longer online but
that I was able to find via archive.org:
http://web.archive.org/web/20081005211317/http://grafi.ii.pw.edu.pl/gbm/x86=
/386.html

Apparently UMOV is only useful in detecting 386DXLV & SXLV chips.
Since I have never heard of a "DXLV" variant and am likely to never
encounter one in my lifetime, I will likely remove the code and forget
about trying to detect AMD parts from Intel.

"Jim Leonard" <mobygamer@nospicedham.gmail.com> wrote in message
news:9eb7d855-080d-45bd-945b-b7ab5e8c59ce@o9g2000yqa.googlegroups.com...
> On Dec 30, 9:50 am, Jim Leonard <mobyga...@nospicedham.gmail.com>
> wrote:
> > Can anyone clarify this? Or, is there a better way to determine if a
> > 386 is an AMD or Intel part?
>
> I think I found my own answer at pages that are no longer online but
> that I was able to find via archive.org: [link]
>
> Apparently UMOV is only useful in detecting 386DXLV & SXLV chips.
> Since I have never heard of a "DXLV" variant and am likely to never
> encounter one in my lifetime, I will likely remove the code and forget
> about trying to detect AMD parts from Intel.

Debbie Wiles processor identification guide via archive.org
http://web.archive.org/web/20040604002243/http://debs.future.easyspace.com/Programming/OS/cpuid.txt

Robert Collins covers UMOV
http://www.rcollins.org/secrets/opcodes/UMOV.html

HTH,


Rod Pemberton

On Dec 30 2011, 8:43=A0am, Jim Leonard <mobyga...@nospicedham.gmail.com>
wrote:
> On Dec 30, 9:50=A0am, Jim Leonard <mobyga...@nospicedham.gmail.com>
> wrote:
>
> > Can anyone clarify this? =A0Or, is there a better way to determine if a
> > 386 is an AMD or Intel part?
>
> I think I found my own answer at pages that are no longer online but
> that I was able to find via archive.org:http://web.archive.org/web/200810=
05211317/http://grafi.ii.pw.edu.pl/g...
>
> Apparently UMOV is only useful in detecting 386DXLV & SXLV chips.
> Since I have never heard of a "DXLV" variant and am likely to never
> encounter one in my lifetime, I will likely remove the code and forget
> about trying to detect AMD parts from Intel.

AFAIK the story was these chips activated Intel's ICE microcode and
called it "SMM", which led to a lawsuit from Intel. In the end AMD
(and IBM which were doing this too) had to remove the microcode and
stop selling these chips.