Re: AMD XP compatibility??

From: Måns Rullgård (mru_at_kth.se)
Date: 06/05/04 

Date: Sat, 05 Jun 2004 17:45:17 +0200

rodsmith@nessus.rodsbooks.com (Rod Smith) writes:

> In article <e08015d.0406050207.7ccff101@posting.google.com>,
> delsherle@hotmail.com (Del) writes:
>>
>> I had RedHat on my old machine and I want to put it on my new AMD
>> Athalon XP 2500 but I can't find an AMD solution that is 32 bit. All
>> I can find is the new 64 bit ones.
>
> AMD's 32-bit Athlon chips (note the spelling -- there's only one "a"),
> including the Athlon XP, are x86 architecture; they run most programs and
> OSs designed for any 32-bit x86 CPU, including Windows and most x86
> versions of Linux. In particular, Red Hat runs just fine on these CPUs.
>
> That said, you *CAN* use CPU-specific features when you compile programs.
> Depending on the precise compiler options you use and which CPUs are
> under discussion, moving a binary from one system to another might not
> work. For instance, if you recompiled your kernel with Pentium 4
> optimizations, I can't guarantee that it'd boot on your Athlon. (Of
> course, non-CPU hardware differences, such as the hard disk controller,
> will also cause problems with that.) If you use the stock binaries,
> including the stock kernel, provided by Red Hat, though, you should be
> able to swap disks and be fine, or re-install the OS from the original
> media.

I've moved a disk with a Gentoo installation optimized for P3 between
a P3 and an AMD Duron machine without any problems. Of course, I
compiled the kernel with appropriate hardware support in advance.

> If you want to fully optimize your software for your hardware, you might
> want to look into Gentoo Linux (http://www.gentoo.org). It lets you set
> compiler options and recompile everything so that it makes the best use of
> the hardware. At least, that's the theory -- how much of an improvement in
> performance it'll make is uncertain, at least in my mind. I suspect the
> difference is pretty small, though.

Most distribution ship binaries that will run on a 486, but are often
scheduled for P2 or P3. I have seen optimization for P4 make programs
up to twice as fast in certain cases, when compared to gcc defaults. 

-- 
Måns Rullgård
mru@kth.se

Relevant Pages

  • Re: 386-586-686
    ... And for that matter, I have an AMD X2 4400+ processor, yet I'm using a ... 386 kernel. ... Would one of the other kernels offer better optimizations ...
    (Ubuntu)
  • Re: kernel mess
    ... Correct semprons were originally Socket A cpu's as a replacement for the Athlon XP etc. AMD ditched the socket A in favour of S754, but disabled the 64 bit instruction on the semprons promoting the AMD64 instead. ... I was using -amd64 while running k8 kernel. ...
    (Debian-User)
  • Re: Code hangs or crashes on Kernel startup (dbau1100)
    ... Is there any debug output at all that we could see? ... is the eboot on your board aware of the different RAM layout? ... We have a new board which is based on the AMD Alchemy ... code, but after it tries to start the kernel, nothing happens. ...
    (microsoft.public.windowsce.platbuilder)
  • Code hangs or crashes on Kernel startup (dbau1100)
    ... We have a new board which is based on the AMD Alchemy ... code, but after it tries to start the kernel, nothing happens. ... putting debug code in the OEMInit function and it never makes it there. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: [opensuse] Distro Question
    ... AMD is submitting patches for these to the Kernel developers. ... AMD 780g chipset and 10.3 and the 11 betas won't even work on it. ...
    (SuSE)