So does this mean that the next ver of Mandrake will
install on my AMDK6-400 without any trouble just like
7.2 ?
If not I will be VERY upset
On Friday 30 March 2001 12:51 am, so spoke David Rankin:
> "Jose M. Sanchez" wrote:
> > To begin with, this is not a Mandrake issue, it's a Linux issue.
> >
> > The K6 can execute i686 code, just not optimally for the processor.
> >
> > With the Kernel it's another matter altogether...
> >
> > The MTRR register differences PREVENTs the i686 compiled kernel to
> > operate on a K6 afaik.
> >
> > When I tried it I got an kernel which would not fully init...
> >
> > That plus the byte alignment optimzations for the 686, make the i586
> > distro better for the K6 in terms of speed and compatibility.
> >
> > It's not a question of making the i686 distro, "properly" identify the
> > k6...
> >
> > LINUX (as directed by Linus) "properly" identifies the K6 characteristics
> > and classifies it as a i585 for optimal operation...
> >
> > There was quite an extensive thread about this some time ago in the
> > kernel forums...
> >
> > It's also doubtful that MANDRAKE will alienate it's prospective clients.
> >
> > I wouldn't doubt that what they'll resort to doing is producing a common
> > distro that installs a i586 kernel with i686 code RPM's... albeit with a
> > slight loss of performance over i586 optimizations or a seperate i586
> > distro, as they've done for the i486.
> >
> > Given the perponderance of AMD's though, I'll bet they will choose the
> > former as it makes good marketing sense...
> >
> > It would be pretty easy to implement in an installer... they already do
> > this for SMP installations anyway...
> >
> > -JMS
> > [EMAIL PROTECTED]
>
> Well put Jose! I hope we see something like this come from Linus. Looking
> down the "time horizon", this issue will probably dissapear in 8 to 12
> months, but in the interim, it just makes good business sense not to raise
> potential obstacles to the further acceptance of Linux.
>
> --
> David Rankin
> Nacogdoches, Texas
