Re: Inconsistent CPUID result

From: babel2017 (babel2017_at_aol.com)
Date: 02/16/04 

Date: Mon, 16 Feb 2004 18:32:27 +0000 (UTC)

"Matt Taylor" <para@tampabay.rr.com> wrote in message news:<gsQXb.173836$fH2.98723@twister.tampabay.rr.com>...
> "Michael Brown" <see@signature.below> wrote in message
> news:MuIXb.45035$9k7.921170@news.xtra.co.nz...
> > babel2017 wrote:
> > > Hi
> > >
> > > I use a machine fingerprint for license control. Have just introduced
> > > CPUID into the mix and it's out for testing. One tester's CPU
> > > suddenly decided to change its name! It was "AMD Athlon(tm)
> > > Processor" and now it's "AMD Athlon(tm) XP 3000+"! Any clues as to
> > > why this happened and whether it's possible to get consistent results
> > > from CPUID?
> >
> > The name string is user-configurable. If there is a BIOS update applied,
> > there is a chance that it will call the chip something different. For
> > example if the user's CPU is now officially supported, it'll possibly go
> > from "AMD Athlon xxxMHz" to "AMD AthlonXP yyyy". Also, if the user
> > overclocks their machine, the name will change as well. It's far better to
> > use the processor ID functions of CPUID, rather than the name string
> > function.
>
> I might also add that on a dual-processor machine there is an even worse
> problem. Oddly enough, my BIOS initializes my first CPU's name string to
> "AMD Athlon(tm) MP 1600+" second CPU's name string to "AMD Athlon(tm)
> Processor" I wrote a quick & dirty cpuid program to demonstrate this:
>
> C:\Documents and Settings\Matt\Desktop>cpuid
> AMD Athlon(tm) MP 1600+
>
> C:\Documents and Settings\Matt\Desktop>cpuid
> AMD Athlon(tm) Processor
>
> Even if the BIOS did set both names correctly, the steppings can be
> different. In fact I know of one person who ran an "unequal" dual-CPU
> machine with Celerons. The first was 300-something MHz, and the second was
> 400 MHz. One must be wary of this case as well. Either take only the first
> processor into account or take every processor into account.
>
> -Matt

Thanks for all the excellent advice. I think I'll try to use the
processor signature instead. But I'm an assembler newbie (not
counting 6502 in AppleII days!) and am not sure how to select which
processor CPUID reports on in a dual-processor setup. Would be
grateful for advice on that.

For others who pass this way, here's a good link I found:

http://www.intel.com/design/Xeon/applnots/24161825.pdf

Relevant Pages

  • Re: Inconsistent CPUID result
    ... my BIOS initializes my first CPU's name string to ... >> processor into account or take every processor into account. ... > Thanks for all the excellent advice. ... The point was that in a dual-processor setup it is not enough to look at the ...
    (comp.lang.asm.x86)
  • Re: SCOAdmin Account Creation
    ... flat file system technology for storing account information. ... use 'webmin' rather than scoadmin: ... Can anybody advice if any of the above processes are holding the ...
    (comp.unix.sco.misc)
  • Re: Exhcnage Email - Internal and External - I know this has been answered but I am still ha
    ... I support 15+ SBS servers, ... Listen to our advice. ... > from their .local account to the outside world. ... I always get brushed off with "Run the CCIEW Wizard!" ...
    (microsoft.public.windows.server.sbs)
  • Re: Changing the Domain Administrator password - advice please
    ... I'd suggest that you revisit your ... design and processes and make sure that there is NO service or process ... It is the same as changing a password on any other user account. ... One piece of advice is to use different accounts to run your services. ...
    (microsoft.public.windows.server.active_directory)
  • Re: accessing urg through outlook express
    ... for me, so I thought, but thanks to helpful advice on another thread I'm ... That's why I asked for her to look below the set of folders as in the second graphic on this URL; ... It shows OE set up with *NO* news server account which should show up below the 'drafts' folder at least. ...
    (uk.rec.gardening)