RPM packages: Athlon vs i686 vs i386

[Jeff Vian]

Don Dixon wrote:

> This link explains it from a 2.6 perspective, but in the FC1 updates 
> directory, there exists an kernel-2.4.22-1.2174.nptl.athlon.rpm and a 
> kernel-2.4.22-1.2174.nptl.i686.rpm.  So which should be applied 
> (keeping in mind what I mentioned earlier)?
>
> Don Dixon
> fedora-list_me_no_like_spam at macrologic.com

I made a similar comment when FC2t1 was released and was told no new 
kernels woud be built for athalon.

The response was as has been said.  The new kernel optimizes for athalon 
automatically.

However, as I understand it, that is the 2.6 kernel.  AFAIK the 2.4 
kernel still gets athalon builds.

>
> Christopher K. Johnson wrote:
>
>> Don Dixon wrote:
>>
>>> Greetings,
>>>
>>> I recently submitted what I thought was a bug to bubzilla concerning 
>>> CPU
>>> detection and FC1.  The bugzilla number is 116941 if anyone is
>>> interested in reading in detail.  The basic problem is that I have an
>>> Athlon XP 1600+ CPU, yet when I boot to the CLI, it says:
>>>
>>> Fedora Core release 1 (Yarrow)
>>> Kernel 2.4.22-1.2174.nptl on an i686
>>>
>>> This exact same hardware, until recently was running RH7.3 and it 
>>> booted
>>> to the CLI identifying itself as athlon.  That was the reason I thought
>>> it was a bug.  On bugzilla, the report has been closed with:
>>>
>>> "athlon is always a i686 for the kernel"
>>>
>>> So, what I am now wondering about is updates.  I need to know which 
>>> level of RPM packages to
>>> load.  For instance with kernels, would I choose 
>>> kernel-2.4.22-1.2174.nptl.athlon.rpm or
>>> kernel-2.4.22-1.2174.nptl.i686.rpm?  And with glibc, do I choose 
>>> glibc-2.3.2-101.4.i686.rpm
>>> or glibc-2.3.2-101.4.i386.rpm.
>>>
>>> Any opinions would be greatly appreciated.
>>>
>>> Don
>>> fedora-list_me_no_like_spam at macrologic.com
>>>
>> Not a problem.  See http://people.redhat.com/%7Earjanv/2.6/readme.txt
>> which includes this note:
>>
>> "I no longer build athlon kernels because in theory the 2.6 kernel 
>> now has a
>> mechanism to automatically runtime patch in these optimisations 
>> during boot"
>>
>> Chris
>>
>
>
>