Re: Linux on Sparcstation4?

From: Paul S. Brown (usenet060103_at_geekstuff.co.uk)
Date: 05/28/04

  • Next message: Goran Larsson: "Re: Linux on Sparcstation4?"
    Date: Fri, 28 May 2004 10:45:43 +0100
    
    

    Casper H.S. *** wrote:

    > "Paul S. Brown" <usenet060103@geekstuff.co.uk> writes:
    >
    > Well, Solaris of any recent for (later than 7) won't run on the IPC/SLC
    > and the like; but Linux is painfully slow on the sun4c hardware because
    > it doesn't deal with pmegs very efficiently (NetBSD is a better choice).
    >

    Agreed. It's not actually too bad on the IPX tho.

    > I don't think it was a late comer; I seem to remember HAL (the original
    > SPARC64 trademark owner) where the SPARC64 line originated did a
    > Solaris 2.4 based SPARC64/OS which was announced/released before Sun
    > shipped Solaris 2.5 on the UltraSPARC hardware which was still a 32 bit
    > OS.
    >
    >>I have just given up trying to figure out what the heck Sun are smoking as
    >>regards their OS strategy as they seem to be on their third or fourth
    >>roadmap in the last few years, having gone as far as getting their IA64
    >>servers WHQL certified which would have been unthinkable just a year ago.
    >
    > Sun doesn't sell/manufacture IA64 servers. I think the windows
    > certification
    > is a no-brainer and a sales enabler. Sun is not EOMing MS Windows so it's
    > not on our OS strategy roadmap.
    >

    OK - AMD64 rather than IA64, but even so - the V20z range.

    What I was pointing at was a 180degree about turn from your board who had,
    up until that point been making "We'll never certify anything with Windows
    and make no moves to provide succour to the enemy" noises, conveniently
    forgetting about the SunPC and SunPCI cards.

    The Sun OS strategy has gone, in the last few years from a pure Solaris play
    to Solaris/JavaOS play back to pure Solaris to joint Solaris/Linux with
    some major personality problems to what appears to be "We'll support
    anything". Just because you're not OEMing Windows doesn't mean that you're
    not going to be expected to provide support for it on your hardware -
    especially for HAL related issues.

    I don't see why the certification is a no brainer as it introduces an
    additional support overhead at a time when Sun is shedding just about
    everybody and their dogs along with doing some very odd things to its
    design groups. Given the ranting we've heard coming from the top for a very
    long time I'd have thought a more coherent strategy for Sun would have been
    to stick with what it knew, not certify with Windows and introduce the
    requirement to support the hardware abstraction elements of Windows and
    carry on that way while reducing costs both in house and to the end user.

    What I'm seeing is a dilution of purpose at a time when technical and sales
    alike are being shed by the thousand and by definition there are fewer
    people trying to support more. That worries me a little.

    P.


  • Next message: Goran Larsson: "Re: Linux on Sparcstation4?"