Suresh Ramasubramanian wrote:
> I should have added "before the 2.5 is released"
Yup... But I expect that soon...
>
> > 2.6 is expected to take as much time because almost all anticipated features are
> > enterprise features like built in clustering support, distributed VM etc. Also
> > If I read in between the line correctly, the 2.6 deadlines will be much more
> > conservatively declared, accounting to the fact that 2.4 was viewed as delay by
>
> They shouldnt fall into the redhat / M$ trap - releasing new version numbers
> just because its all the rage.
No. It's other way round. Since most of the marketing people don't understand
time scale of developers, expand the time scale itself. Tell them that it's
going to be 4 months and release it in 3 months. That will give them some
pleasure.
I don't think kernel team will give any damn to any time table. There can always
be new version to accommodate new change/feature. And nobody is prohibited to
release their own patch against the kernel.
>
> > More than kernel, interesting thing is going to be Intel/AMD aka
> > itanium/sladgehammer war. My guess is intel will be on much more backfoot
> > compared to P4/Athlon war. Linux will be only OS available, ready to run on it.
> > That will be interesting scenario.
>
> Nah, Intel has more than enough muscle to tide this over, IMHO at least.
> But then, this was the case with (say) the PDP series before the PC came on
> the stage :)
If AMD overcomes with their production problem, release a good chipset (in
volume) of their own and introduce a good compiler/compiler optimisations to gcc
types, then Intel has much to sweat about. Of course inertia is one thing Intel
has in ampule and that's difficult to compete with.
I dunno but AMD spefic optimisations compared to P* specific ones, are always
not good enough to me. By some benchmarks, optimisations made P4/1.7GHz to pull
ahead of Athlon 1.33GHz.(Yesterdays news. Do't remember. Mostly AnandTech/Tom's
Hardware). Athlon didn't have any optimisation to offer. Thst's bad for AMD if
you ask me.
Itanium performance is going to heavily depend upon compiler due to EPIC(And
since M$ takes 2 years to come up with processor specific optimisations in
compilers, it's expected to screw it excellently) while sledgehammer's is not.
AMD should really come up with their own compiler. At least contribute to GCC
generously. That would boost their market position.
And I think it's time to think as AMD and Intel are separate architectures and
start shipping separate products for them. Even linux kernel may witness some
small fragmentation over architectures. (All my guess but I count on it...)
Shridhar
----------------------------------------------
The mailing list archives are available at
http://lists.linux-india.org/cgi-bin/wilma/linux-india-help
