Future of SGI? - SGI

This is a discussion on Future of SGI? - SGI ; I read somewhere that SGI hired "turnaround artists" (sic) last year, in order to save the company. In the meantime, SGI stock is selling for peanuts. Does anybody know what is really going on at SGI? What's the future direction ...

+ Reply to Thread
Page 1 of 3 1 2 3 LastLast
Results 1 to 20 of 46

Thread: Future of SGI?

  1. Future of SGI?

    I read somewhere that SGI hired "turnaround artists" (sic) last
    year, in order to save the company. In the meantime, SGI
    stock is selling for peanuts. Does anybody know what is
    really going on at SGI? What's the future direction for SGI?

    I read the "SGI Strategy" whitepaper (PDF, http://tinyurl.com/7zguk)
    however will this really save SGI? And make it a viable
    and flourishing company again? Thank you for any
    comments and insights.



  2. Re: Future of SGI?

    I don't know all the details but personally I think SGI need to sack
    their directors and get some forward-thinking people in. From what I
    can gather they've missed the boat as far as beating the competition
    goes (which is now employing SGI's ex-brightest by the looks of
    things), and now need to get back on track as a viable going concern.
    I think also the market has changed, so perhaps there's just not enough
    of a market left for the kind of machines SGI make. Mind you,
    high-performance sports car companies have only survived by being
    bought out by the larger companies who traditionally made cheap
    ordinary units, maybe SGI need to be bought out and kept running as a
    niche supplier.


  3. Re: Future of SGI?

    wrote


    > I think also the market has changed, so perhaps there's just not enough
    > of a market left for the kind of machines SGI make. Mind you,
    > high-performance sports car companies have only survived by being
    > bought out by the larger companies who traditionally made cheap
    > ordinary units, maybe SGI need to be bought out and kept running as a
    > niche supplier.


    .... or SGI should become a larger company itself, starting
    to sell ordinary units as well.

    For example, why couldn't SGI
    create a graphic card that is superior to NVidia's and ATI's
    offerings? After all, they invented OpenGL and computer
    graphics, so to speak. I'm sure that kids, playing all those
    OpenGL-based games, would buy such a thing
    in a heartbeat. It would probably become some kind
    of a status-symbol. And I'm sure SGI, with all their technical
    expertise, could make such a card.

    Or, why can't SGI develop small, cheap but high quality
    servers? (Yes they already have such a thing but they're
    still too expensive IMHO. The goal should be to undersell
    HP and Dell.)

    Finally, why couldn't SGI compile their own, certified Linux distribution,
    and sell it together with their hardware? "SGI Linux", a Red Hat killer.
    I would buy it in a heartbeat. (Only if it follows high SGI
    standards, of course.) Generic Linux is fine but IMHO
    SGI need their own, certified, and fully compatible distribution.



    Weow. It's hard to watch SGI sinking, and sinking, and sinking.
    I only they would do something, anything, fast.

    It's so
    sad to watch the best company in the history of IT go under!



  4. Re: Future of SGI?

    In article , "kolaric" wrote:

    : For example, why couldn't SGI
    : create a graphic card that is superior to NVidia's and ATI's
    : offerings?

    There's two basic technologies SGI fell behind on: Graphics and CPUs.

    The reason for this is the fact that other companies (Intel, AMD, nVidia, ATI)
    are operating on razor thin margins on huge volumes, and pushing their product
    release schedules right to the breaking point.

    Basically, if a company is not already on the treadmill at this point, they have
    essentially no chance of ever catching up without a very huge and very risky
    capital investment.

    The only company lately that has tried to break into or re-enter the graphics
    card market is VIA/S3 with their Chrome series of graphics cards. It's a pretty
    safe bet that you've never seen one of these cards, and even more probable that
    you've never considered buying one.

    It's a chicken and egg deal, you can't support the low margins without the
    volume, and you can't get the volume without the low margins.

    This is on top of the fact that SGI long ago lost most of their graphics
    engineers to nVidia, and most of their CPU engineers to Intel.


    SGI's workstation market is gone and it's never coming back.

    It remains to be seen if SGI's big iron Itanic market is sustainable, but in the
    mean time they're competing relatively well with IBM and other supercomputer
    makers, and they still have a few tricks in the pipeline such as integrating
    vector processors and FPGAs directly into their supercomputer systems.

    But only time will tell.


    Cheers - Tony 'Nicoya' Mantler

    --
    Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@ubb.ca
    -- http://nicoya.feline.pp.se/ -- http://www.ubb.ca/ --

  5. Re: Future of SGI?


    > This is on top of the fact that SGI long ago lost most of their graphics
    > engineers to nVidia, and most of their CPU engineers to Intel.



    I'm convinced that if SGI is to be saved,
    it will have to be done through computer
    graphics technology.

    SGI is foremostly a computer graphics company. Yes
    it produces great supercomputers. However that's irrelevant
    in this context. If SGI really wants to thrive, it needs
    to abandon its "niche" supplier status selling perhaps
    1 supercomputer per month, and seek to enter wider markets.



    > SGI's workstation market is gone and it's never
    > coming back.



    > It remains to be seen if SGI's big iron Itanic market is sustainable, but
    > in the
    > mean time they're competing relatively well with IBM and other
    > supercomputer
    > makers, and they still have a few tricks in the pipeline such as
    > integrating
    > vector processors and FPGAs directly into their supercomputer systems.


    No, big iron Itanic market is definitely not sustainable.
    I mean, that's obvious, isn't it? SGI revenues and profits
    fall by every passing year. What good have
    supercomputers (or any other technology except computer
    graphics) done to SGI? None.

    Graphics cards: I don't believe it's a chicken and egg deal.
    Remember the Voodoo 3dfx SLI craze?
    SGI folks just need to sit down, set objectives and make
    a card that's 10x better than anything else on the market.
    They can charge a premium on such a card. I'm sure
    everybody and his brother would like to have an SGI
    card, 10x better than anything, in his PC.



  6. Re: Future of SGI?

    In article , "kolaric" wrote:

    : SGI folks just need to sit down, set objectives and make
    : a card that's 10x better than anything else on the market.
    : They can charge a premium on such a card. I'm sure
    : everybody and his brother would like to have an SGI
    : card, 10x better than anything, in his PC.

    There's no magic pixie dust at SGI. The days when a company could come out of
    left field with a product 10x better than the competition are long over.
    Everybody is hitting their heads against the hard limits of math and physics and
    only making incremental progress, and there's no easy way around it.

    And even if they did come up with some miracle product, they'd have to do it
    again 12-18 months down the line, 'cause everyone else will have caught up by
    then. One thing SGI was never able to do was ramp up to a fast paced development
    cycle that could consistently deliver an improved product on a quasi-yearly
    timeframe. Even if they could have, they never had the volume to pay for such a
    massive amount of engineering.

    So, you can dream all you like about the second coming, but it won't happen in
    the form of a cheap graphics card or workstation.


    Cheers - Tony 'Nicoya' Mantler

    --
    Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@ubb.ca
    -- http://nicoya.feline.pp.se/ -- http://www.ubb.ca/ --

  7. Re: Future of SGI?

    Le Sat, 28 Jan 2006 01:18:08 +0000, Tony 'Nicoya' Mantler a écrit*:

    >
    > It remains to be seen if SGI's big iron Itanic market is sustainable


    It looks more and more like they bet on the bad horse... Itanium is more
    itanic every year, while Opteron soars. Did you see the numbers ? In 2005,
    around 2000 Itanium systems were shipped WORLDWIDE. Like in 2004,
    and like in 2003, and 2002. Itanium is going exactly nowhere; by chance
    Intel seems to keep it running but how long will that last ?

    --
    Quidquid latine dictum sit, altum sonatur


  8. Re: Future of SGI?

    Le Sat, 28 Jan 2006 16:01:27 +0100, Michael Laajanen a écrit*:

    >
    > I read on /. that Intel and others will spend $10billion on Itanium in the
    > comming years, if it will help Itanium who knows?


    Lip service, my friend. They delay or cancel more and more itanium
    developments : the new chipset (supporting both Itanium and Xeon) was
    delayed (perhaps forever...), the x86 compatibility layer in Itanium was
    removed, Microsoft ceased development of Itanium ports for several apps,
    and the hardware development is getting slower and slower.

    Dell, IBM and others all left the sinking ship. There are actually only 3
    Itanium providers remaining : SGI, HP and Bull. HP of course is too
    involved in Itanium, and Bull and SGI are both niche players, in equally
    very bad shape. If HP gets in trouble someday, be sure the Itanium
    will die right away.

    --
    Je suis riche des biens dont je sais me passer.
    Louis-Jean-Baptiste Etienne Vigée.


  9. Re: Future of SGI?


    > You're assuming nVidia and ATI don't have similar levels of expertise,
    > and that out-doing those companies would be easy. That would be in
    > error. Who do you think poached SGI graphics experts?


    Yes but there are others graphics experts in the world, aren't there?
    Talent is limited but nVidia can't hire them all, can they?



    >>+ still too expensive IMHO. The goal should be to undersell
    >>+ HP and Dell.)

    >
    > They've tried that. It didn't work so good the first time. What makes
    > you think it'll go better the second time around?



    Something has to be done. Simple as that.




    >>+ Finally, why couldn't SGI compile their own, certified Linux
    >>distribution,
    >>+ and sell it together with their hardware? "SGI Linux", a Red Hat
    >>+ killer.

    >
    > And piss off the IRIX loyalists further? oy.


    In the great scheme of things, it's Linux vs. Windows now.
    IRIX is an advanced OS but I'm afraid it lost momentum,
    along with other similar unices.

    SGI has to decide. You can't support two OSes in your house.
    It drains resources, blurs focus.





    >>+ I would buy it in a heartbeat. (Only if it follows high SGI
    >>+ standards, of course.) Generic Linux is fine but IMHO
    >>+ SGI need their own, certified, and fully compatible distribution.

    >
    > And how much can they afford to spend on this? my guess is that this
    > would be yet another sinkhole for their few remaining dollars. And
    > which distro should they use as their base? I'm thinking that
    > Slackware would be best, as it's neither Red Hat nor Debian. They
    > might be able to work a deal with Gentoo, but I'm not so sure that the
    > Gentoo crowd would be all that exicited about such a deal. I could be
    > very wrong, tho.
    >
    > Do they have enough people left to do something like that completely
    > in-house?



    Well, I just know I want SGI Linux.

    I want to run SGI Linux on a SGI computer featuring a mighty
    SGI OpenGL graphics card, using 3D-peripheral devices also
    made by SGI.

    An "SGI computer" could be a PC, but it has to be assembled
    at SGI, passed all their tests of quality.






    >>+ Weow. It's hard to watch SGI sinking, and sinking, and sinking.
    >>+ I only they would do something, anything, fast.

    >
    > I agree, they're circling the drain. I agree that if they're going to
    > attempt to right the ship, the should do so with all possible
    > speed. But they also shouldn't do anything rash. They need to figure
    > out a strategy that has the highest probability of success.



    That whitepaper that I mentioned in my OT certainly
    doesn't describe such a strategy. It's a sad, apologetic document,
    leading to nowhere.



  10. Re: Future of SGI?


    > : SGI folks just need to sit down, set objectives and make
    > : a card that's 10x better than anything else on the market.
    > : They can charge a premium on such a card. I'm sure
    > : everybody and his brother would like to have an SGI
    > : card, 10x better than anything, in his PC.
    >
    > There's no magic pixie dust at SGI. The days when a company could come out
    > of
    > left field with a product 10x better than the competition are long over.
    > Everybody is hitting their heads against the hard limits of math and
    > physics and
    > only making incremental progress, and there's no easy way around it.


    Point taken. Yes, there are limits. However if you sit down, set objectives,
    you can do it. You can find a way. Try to see the problem from another
    viewpoint. And still make a 10x better product.




    > And even if they did come up with some miracle product, they'd have to do
    > it
    > again 12-18 months down the line, 'cause everyone else will have caught up
    > by
    > then. One thing SGI was never able to do was ramp up to a fast paced
    > development
    > cycle that could consistently deliver an improved product on a
    > quasi-yearly
    > timeframe. Even if they could have, they never had the volume to pay for
    > such a
    > massive amount of engineering.


    Fast development cycle - well it's just a matter of organization.
    Again, you have to set objectives. I don't think SGI is intrisincally
    "slow"
    or whatever. If they were slow in the past that means they have
    to change.


    > So, you can dream all you like about the second coming, but it won't
    > happen in
    > the form of a cheap graphics card or workstation.


    Yes, I'm well aware of the fact that the current situation at SGI
    is extremely bleak. However hope dies last. I'd simply like to see
    the best company in the business, on the throne again. This desire
    of mine is understandable, isn't it?




    I see it like this:

    0) keep selling super-computers and visualization solutions
    to universities, army, companies, etc., to stay afloat

    1) use this "given" time to design a
    great OpenGL card for PCs

    2) assemble a SGI Linux distribution, outphase IRIX

    3) later on, with profits from OpenGL cards, launch
    a "SGI PC" - a high-quality PC with a moderate price,
    for a) gamers, b) visualizers, c) designers. In short,
    for gamers and creative folks. Launch this machine with
    a great collection of software.

    4) design and sell other 3D peripheral devices, like 3D mice,
    headsets, VR helmets.

    5) continue working on state-of-the-art 3D software,
    like Performer.

    6) Lauch a free, advanced 3D engine API/SDK for
    gamers, running on Linux.

    7) merge Performer with Inventor (i.e. continue work
    on what was supposed to become OpenGL++/Fahrenheit);
    continue work on OpenGL

    8) of course, continue designing great supercomputers.



  11. Re: Future of SGI?

    >
    > I want to run SGI Linux on a SGI computer featuring a mighty
    > SGI OpenGL graphics card, using 3D-peripheral devices also
    > made by SGI.


    Well, this is not at all what's being offered by sgi...
    You can get a Suse or RedHat Linux with some sgi stuff on top of it; and
    the graphics are ATI... Making yet another Linux distribution and
    starting from scratch on graphics hardware makes no business sense
    whatsoever, especially for a company as deep in the red as sgi...

    Just out of curiosity, how much are you willing to pay for such a
    workstation? Would your price range for a Linux sgi workstation be
    comparable to the UNIX workstations or would you rather expect something
    similar to a Windows workstation? I'm betting the latter... There is a
    French expression that fits perfectly here: "vouloir le beurre et
    l'argent du beurre"...

    I'll let Emmanuel explain that one ;-)

  12. Re: Future of SGI?

    HI,

    Emmanuel Florac wrote:
    > Le Sat, 28 Jan 2006 16:01:27 +0100, Michael Laajanen a écrit :
    >
    >
    >>I read on /. that Intel and others will spend $10billion on Itanium in the
    >>comming years, if it will help Itanium who knows?

    >
    >
    > Lip service, my friend. They delay or cancel more and more itanium
    > developments : the new chipset (supporting both Itanium and Xeon) was
    > delayed (perhaps forever...), the x86 compatibility layer in Itanium was
    > removed, Microsoft ceased development of Itanium ports for several apps,
    > and the hardware development is getting slower and slower.
    >
    > Dell, IBM and others all left the sinking ship. There are actually only 3
    > Itanium providers remaining : SGI, HP and Bull. HP of course is too
    > involved in Itanium, and Bull and SGI are both niche players, in equally
    > very bad shape. If HP gets in trouble someday, be sure the Itanium
    > will die right away.
    >

    I will for sure not miss it. They killed Alpha and HPPA which had a
    market, I actually worked designing cPCI HW for EV7(EV67 during the dev)
    it was a mighty one!

    Now we have to hope for the Fujitsu guys to create a great Sparc so we
    still have a few high end CPUs in this world.

    /michael



  13. Re: Future of SGI?


    >> I want to run SGI Linux on a SGI computer featuring a mighty
    >> SGI OpenGL graphics card, using 3D-peripheral devices also
    >> made by SGI.

    >
    > Well, this is not at all what's being offered by sgi...


    Of course it isn't offered. (At the moment.)



    > You can get a Suse or RedHat Linux with some sgi stuff on top of it; and
    > the graphics are ATI... Making yet another Linux distribution and starting
    > from scratch on graphics hardware makes no business sense whatsoever,
    > especially for a company as deep in the red as sgi...



    SGI wouldn't start from scratch. SGI still sell great visualization
    hardware systems, and are makers of OpenGL, Inventor, Performer...
    I would hardly call this "starting from scratch". SGI has
    absolutely all the right preconditions to enter commodity
    graphics card market. If not SGI who then?

    "Yet another Linux distribution" wouldn't be just that.
    It would be **SGI Linux**, Linux prepared from the best
    company that has ever existed in the business. Linux carefully
    prepared, tested and augmented to run on a quality
    SGI PC/WS and servers.

    If not Linux, then SGI should continue developing IRIX.
    But please decide. You can't sit on two chairs!
    It simply confuses buyers, and makes them wonder
    whether SGI simply isn't able to focus and deliver.



    > Just out of curiosity, how much are you willing to pay for such a
    > workstation? Would your price range for a Linux sgi workstation be
    > comparable to the UNIX workstations or would you rather expect something
    > similar to a Windows workstation? I'm betting the latter...


    Something in-between. Take the most expensive
    Dell or HP workstation-like PC. I would be ready to
    pay this + perhaps 10-20% more. +30%
    would already be too expensive.


    > There is a French expression that fits perfectly here: "vouloir le beurre
    > et l'argent du beurre"...


    Have a cake and eat it too?





  14. Re: Future of SGI?

    Tony 'Nicoya' Mantler wrote:
    >In article , "kolaric" wrote:
    >
    >: SGI folks just need to sit down, set objectives and make a card that's
    >: 10x better than anything else on the market.


    Citing the overwhelming success of the first Voodoo 3D graphics cards in
    the peecee market is a very misleading example. In 1995, 3D vector graphics
    was already a very mature technology, with ~20 years of worldwide research
    and ~10 years of multivendor CAD/CAM workstation market preceding it.

    But because there had been _nothing_ remotely like affordable hardware
    accelerated vector graphics on the peecee market before, made the Voodoo
    cards seem like technomagic.

    And remember, the (pseudo-)high-tech 3D graphics market crashed in ~2000.
    Nowadays virtually every low-end commodity graphics chipset supports 3D
    graphics. The high-end Ati and Nvidia cards are primarily aimed at the 3D
    gamers niche. 3D Gamers are admittedly a highly vocal group, but they are a
    very small minority, about less than 1-5 percent of the graphics market..

    I think it would be madness for a small high-tech engineering firm like SGI
    to enter the high volume commodity graphics market. Matrox bowed out of it
    a few years back because they could cope.

    >There's no magic pixie dust at SGI. The days when a company could come out
    >of left field with a product 10x better than the competition are long over.


    A company that would concentrate itself on high reliable _systems_ with much
    of the hardware and software integrated and hidden could make it, when|if
    reliable systems start to become popular.

    scs

  15. Re: Future of SGI?

    Emmanuel Florac wrote:
    >Le Sat, 28 Jan 2006 01:18:08 +0000, Tony 'Nicoya' Mantler a écritÂ*:
    >> It remains to be seen if SGI's big iron Itanic market is sustainable

    >
    >It looks more and more like they bet on the bad horse... Itanium is more
    >itanic every year, while Opteron soars.


    They bet on two dead horses at the same time. With choosing Linux they
    chose to throw away most if not all of their multi-CPU support. Now they
    have to shoehorn that back into an 'architecture' that isn't fit for it.
    What a waste of effort.

    scs

  16. Re: Future of SGI?

    > I think it would be madness for a small high-tech
    > engineering firm like SGI to enter the high volume
    > commodity graphics market. Matrox bowed out of it
    > a few years back because they could cope.


    Perhaps you're right.



    > A company that would concentrate itself on high reliable
    > _systems_ with much of the hardware and software integrated
    > and hidden could make it, when|if reliable systems start
    > to become popular.


    I agree. However, SGI could take this Dell-like, just in time
    system integration even further, for example by
    shipping machines with its own well-tested and well-integrated
    "SGI Linux" distribution, loaded with all kinds of 3D applications.
    I would certainly buy it.



  17. Re: Future of SGI?

    >>It looks more and more like they bet on the bad horse... Itanium is more
    >>itanic every year, while Opteron soars.

    >
    > They bet on two dead horses at the same time. With choosing Linux they
    > chose to throw away most if not all of their multi-CPU support. Now they
    > have to shoehorn that back into an 'architecture' that isn't fit for it.
    > What a waste of effort.


    Yes.
    It would be interesting to hear from SGI management
    on what was the rationale behind these decisions.



  18. Re: Future of SGI?

    In article <4448odF9dc7U2@individual.net>,
    S.C.Sprong wrote:
    >They bet on two dead horses at the same time. With choosing Linux they
    >chose to throw away most if not all of their multi-CPU support. Now they
    >have to shoehorn that back into an 'architecture' that isn't fit for it.
    >What a waste of effort.


    The results from what has been successful in the marketplace suggest
    that the single-system-image approach has turned out to be a niche
    market compared to fast & cheap independant systems with message
    passing.

  19. Re: Future of SGI?

    kolaric wrote:
    >
    > Hi,
    >
    > You're talking about a full-blown workstation here.
    > Such machines are usually sold to academia, government
    > and industry.


    Well, what else is there? Ah, right, the "hobbyist"... So, are you
    saying sgi to survive must become another Dell?!?!?

    >
    > For SGI to get there (that is, be able to start making
    > and selling such workstations again), IMHO they first have
    > to get back to basics and be able to make great desktop
    > machines for a wider audience - gamers,
    > and professionals working in design, multimedia, visualization.


    Just what I thought: gamers... :-D

    >
    > Competitor in this space would be for example Dell XPS 600.
    > However, the idea is to make something that is better
    > than this, for a slightly (10-20%) higher price. The objective
    > would to be on the bleeding edge of computer graphics.
    > I'm sure that the targeted customers (gamers, creatives) will
    > take notice. Especially gamers - there is a lot of customers
    > out there who will pay premium for such a machine.
    >
    >


    sgi is not at all in that market, sgi makes most of its revenue from
    support, not selling hardware... There is close to no margin in "gamers"
    hardware...

  20. Re: Future of SGI?

    Le Sun, 29 Jan 2006 16:27:57 +0000, S.C.Sprong a écrit*:

    >
    > They bet on two dead horses at the same time. With choosing Linux they
    > chose to throw away most if not all of their multi-CPU support.


    This is grossly exwagerated. There are 512-cpus linux nodes nowadays.

    > Now they
    > have to shoehorn that back into an 'architecture' that isn't fit for it.
    > What a waste of effort.


    Actually Linux was probably a sane choice, the big mistake remains the
    Itanium.

    --
    De longs désirs, une longue admiration sans espérance, voilà le moyen
    d'adorer les femmes, et de rendre l'amour une passion délicieuse!
    N. Rétif de la Bretonne.


+ Reply to Thread
Page 1 of 3 1 2 3 LastLast