SGI to cut 600 more jobs -SJ Mercury

[mental]

this month's issue of Business 2.0 has a very interesting article specifically about CRAY and it's tenious relationship with SGI. it made me wonder about Alias' relationship with SGI.

[SheepFactory]

wow , looks like the Sgi stock is cheaper than a pack of gum.

0.98 cents as of today :rolleyes:

[Cararan]

Originally posted by Sheep Factory
wow , looks like the Sgi stock is cheaper than a pack of gum.

0.98 cents as of today :rolleyes:


.98 cent whoa thats crazy i thought it would be higher than that whats going on??

[wgreenlee1]

Crap...
Does this mean SGI is going to be bankrupt soon?

Alias might be sold?

:shrug:

[Peter Reynolds]

50,000 CG talk members.

How much do we all need to chip in to buy Alias?

[DeathBrain]

Originally posted by wgreenlee1
Does this mean SGI is going to be bankrupt soon?

It could be...PC's and Mac's rules the market these days :)
Like Itanuim, Opteron, G5, Dual things, etc...and lots of brandnew Extreeme Box :)

[SheepFactory]

nah , If you guys go and check , SGI's stock is bouncing between 0.50 <=> 1.50 for the last two years.

they have big goverment contracts that'll keep em going until the goverment\military starts buying intel\amd systems from dell or something :)

[Pixelmaestro]

Example: ASCI White

http://www.llnl.gov/asci/platforms/white/hardware/

Google will be leasing the former SGI property.

SGI might survive as a real estate company.

[Titus]

Originally posted by Peter Reynolds
50,000 CG talk members.

How much do we all need to chip in to buy Alias?

$1,000 per member.

[Thalaxis]

If we time it correctly, we'll be able to run a CGTalk buyout of SGI
for $8/person :)

[trthing]

That would be brilliant!!
Count me in!!

:applause:

[richcz3]

Does anyone remember where nVidia's engineers came from?

After nVidias success SGI sued...

"We believe this action is motivated by the loss of some of the best 3D architects in the world to NVIDIA Corporation," said Jen-Hsun Huang, president and CEO of NVIDIA.

"These people had no involvement in the design of NVIDIA's current RIVA graphics processors which are the subject of the alleged patent infringement. NVIDIA is prepared to vigorously defend itself."

Then they kissed, made up... and went to bed..
http://www.businesswire.com/webbox/bw.081099/192220199.htm

Wonder where SGI's fortunes will head now?

richcz3

[Thalaxis]

Originally posted by richcz3
Does anyone remember where nVidia's engineers came from?


I remember that quite well, yes.

I wonder what their next workstation will be like? Another $20k
monstrosity with 1-4 souped up embedded processors?


Wonder where SGI's fortunes will head now?


What fortunes? SGI has fortunes? :)

[beaker]

Originally posted by Pixelmaestro
Example: ASCI White

http://www.llnl.gov/asci/platforms/white/hardware/

Google will be leasing the former SGI property.

SGI might survive as a real estate company.
Unfortunatly SGI sold off most of it's land and made a deal to leasing that land back from the company it sold it too for just the buildings they had left over for operations. The land itself was worth millions(270million according to a website I found). I went to their campus back in 1998 for training and it was huge. Very beautifull architecture and buildings with lots of open land and trees with trails and such. Sad that they had to part with most of it.

[CADster]

Originally posted by wgreenlee1
Does this mean SGI is going to be bankrupt soon?


they only did it to themselves.

good damn riddence IMO.

[Thalaxis]

Originally posted by CADster
they only did it to themselves.


I definitely agree with you there.

[MacRonin]

Looks like it could be time for Apple to step in and take Alias|wavef, er, I mean Alias off of SGIs collective hands...

Mmmm, Apple Maya Unlimited...!

Flame on!

;^p

[onscreen]

Well, not sure and i am really got shocked when i see SGI keep pumping out data storage servers and stuff. The workstation is still wooing me when i fell in love with the SGI Fuel. Ya it is expensive... really expensive.

Alson, i dont think Apple will but Alias. Coz in thier mind there is no way for them to improve thier computers to be a renderbox and stuff. IF those high-end gfx companies started to port thier cards into Apple line.. wow, there we are, a new breed of 3dbox and renderbox with style... just like Alienware and Boxx

[swami3d]

I truly weep for a company with so much potential that just seemed to keep flushing $$ down the toilet. Case in point:

In 1998 or so when SGI spent $170 million on a new building in Mountain View, they created an architectural marvel for their new headquarters. It was a beautiful building, but there were a few problems.....

1. There was a section of the building which was all glass, and many of the executives had their offices in this area. The glass sections though, were angled, and this literally caused people who had offices in this area vertigo. Some people were so nauseous they had to turn their desks away from the glass so they wouldn't ralph.

2. Air conditioning ducts were routinely blowing people out of their cubes, quite literally. Many people couldn't even work in their designated spaces because it was like an icebox.

3. The wiring for networks, phones, etc. was set up to be a design element of the building, and if you looked up you'd see it all in these metal trays that ran throughout the building. As time went on, these trays got really screwed up, and wires started protuding from all areas, and got really unwieldy.

And that was just their new building, forget about their actual products. Anyone remember the Visual Workstation? Best thing about it was the long sleeved shirt I got at the launch. :)

Chris

Oh, and screw Rick Belluzzo. He left SGI for Microsoft, and I am CONVINCED he was sent to SGI to try to bankrupt it. It's the only logical explanation, considering what he did when there.

[Thalaxis]

Originally posted by swami3d

Oh, and screw Rick Belluzzo. He left SGI for Microsoft, and I am CONVINCED he was sent to SGI to try to bankrupt it. It's the only logical explanation, considering what he did when there.

You are not the first person I have heard to put forth that theory.

The only counter-argument is Mike Capellas; after running
Compaq into the ground, Comcast (that was it, right?) actually
HIRED him. It's a weak counter-argument indeed, eh? ;)

In all honesty, I agree with you.

It does raise the question of motivation, though; the only thing I
can think of is technology; MS and SGI have not been direct
competitors for much of anything, so the only thing I can think of
is that MS wants some of SGI's technology... my guess, based on
their relatively recent acquisition of some SGI graphics-related IP
is that they plan to use SGI 3D-rendering technology in Longhorn,
and therefore in also in the corresponding version of DirectX, in
light of the fact that Longhorn's UI is purportedly going to make
use of a multi-threaded, 3D-rendered UI.

[swami3d]

SGI gets bright idea to enter NT workstation market. Cool.

Announces ambitious machine at extremely good price point, promises fastest 3d around and serial digital video i/o standard. Cool.

Delivers the biggest piece of crap machine that DOES NOT work ever to come off an assembly line, and to top it all off, has CUSTOM NT code written by MS, which makes NT even MORE unstable than that "blue screen of death" piece of shiite it already was.

THEN, Belluzzo (sp?) dives out of SGI to become the VP of Internet tech at MS.

If I were an SGI shareholder (Whew, glad I don't play the market) I would sue the crap out of Rick. He is, in my opinion, the worst thing to ever happen to SGI.

I even wrote an open letter to him in 3D mag about how NOT to screw the Visual PC up, and he did the exact opposite of EVERYthing I mentioned. Prick.

chris

[Thalaxis]

Originally posted by swami3d
SGI gets bright idea to enter NT workstation market. Cool.

Announces ambitious machine at extremely good price point, promises fastest 3d around and serial digital video i/o standard. Cool.

Delivers the biggest piece of crap machine that DOES NOT work ever to come off an assembly line, and to top it all off, has CUSTOM NT code written by MS, which makes NT even MORE unstable than that "blue screen of death" piece of shiite it already was.


Don't forget that the VPC was over a year late, and by the time it
was announced, Intergraph was gearing up to launch their next-
generation WildCat hardware... and on top of that, the VPC's
graphics performance was only around 1/2 that of the WildCat
that was already on the market -- and cost LESS.

Their biggest blunder, IMO, was that the graphics chip was
integrated, and hence not upgradable. It guaranteed that their
machine was uncompetitive, and that improving it required a new
motherboard.

Translation: suicide. You don't compete by offering less for more,
and making up for it with fewer options, less upgradability, and
less stability.

[swami3d]

In one of my back page Op-eds I wrote an open letter to Rick Belluzzo. Does anyone have that copy of 3D Design? I'd like to have a copy and can't find it. I did find this tho, interesting in lite of this discussion to look back in time a bit.

3D DIRECT • February 12, 2000

Eternal Truth of the Month
Why SGI Just Won't Die, and Other Metaphysical Stuff
by Swami Rendalotsa

The New Year and the official turn of the Millennium have had this soothsayer feeling a bit nostalgic, of late. Although the past may be looked upon fondly in many respects, there are always regrets. Always "what ifs." Would it have been different if only I had done this? Done that? Ah, as the saying goes, "Hindsight is 20/20." Your past can be crystal clear at times. I like looking back occasionally, and today my third eye focuses on SGI (www.sgi.com), the company formerly known as Silicon Graphics, Inc.

Ah, and what a varied and complex history to relive. Silicon Graphics Inc., founded in 1982 by Jim Clark, was once the only game in town for the computer graphics industry. This isn't to say there weren't other types of computers that performed graphics tasks such as the Commodore Amiga, or even the short lived Tandy Color Computer. But if you were truly serious about computer graphics in the 80's and into the early 90's, Silicon Graphics, Inc. was the only way to go. And you'd better have lots of gold pieces in your little change purse, because going "SGI" was a major investment.

Silicon Graphics did (and still does) have other types of customers besides the Hollywood set. But the entertainment industry really pushed the envelope in CG, and plus, Hollywood made great grist for the public relations' mill. Stories circulated about the latest big-budget Hollywood film, and how Silicon Graphics was key to the effects work. It was a perfect synergy. With Silicon Graphics as the only maker of computer graphics workstations for high-end visual effects and 3D software, they could charge any price they wanted. And they did. Anyone who tried back then to buy an external CD-ROM drive for their SGI box found it could cost more than $1300.

There were two large influences on the proliferation of 3D on the desktop, and SGI can't take credit for either of them. The first was a little company in Topeka, Kansas, named NewTek (www.newtek.com). At the time (mid to late 1980's), NewTek was selling its Video Toaster video editing system, which was delivered on an Amiga 2000 computer. The unit could not only do things like edit video and provide transitions, users also found that it contained a little extra that seemed to be thrown in as almost an afterthought—a full fledged 3D program, which would become LightWave 3D. LightWave is perhaps the 3D software package most responsible for where we are in 3D on the desktop today, as its rapid growth, increased power, and functionality made the prices of the high-end 3D software on SGIs come down dramatically.

This started a price war in the 3D software space that, to this day, has not abated. The other thing that brought Silicon Graphics down as the king of the hill in CG was the NT workstation. At SIGGRAPH '94 (www.siggraph.org) in Orlando, FL, Intergraph (www.intergraph.com) was the only NT workstation vendor. They occupied a nondescript, and virtually ignored, booth in the back of the conference hall. Although in use for a while by architects and industrial designers using Intergraph MicroStation software, the CG community at SIGGRAPH largely laughed them off as toys, or worse.

Fast-forward six years in time, and the NT workstation is the de facto standard in the world of 3D graphics. Even many Mac diehards (myself included) will agree that powerful 3D workstations in the PC world offer more choice, and faster graphics. The gap is narrowing, but the NT machines have an extremely strong lead. Now, the graphics card in my off-the-shelf clone PC has more power than an entire SGI workstation from five years ago. Silicon Graphics boxes that had cost $50,000 are now being turned into speaker enclosures and cappuccino makers. SGI really missed the boat on the whole PC graphics revolution, and the tides are threatening to wash the company to sea for good.

So what can SGI do to turn it around? Apple did it, so why can't SGI? In a bold move to corner more of the high-end NT workstation market, SGI recently bought Intergraph's Computer Systems division. Noted for their "Z" series of NT workstations, SGI made the announcement at SIGGRAPH 2000 on the morning the show floor opened. Seen around the floor were the big black behemoths we've come to expect, but wearing the SGI logo! An editor at a tech pub likened it to the Titanic, with SGI being the doomed ship. Except, when they see the iceberg, they turn, only to purposely smack into another ship and sink both of them.

SGI then took their sweet time putting any information online. More than a month passed before these machines even went up for sale on the SGI site.
SGI has also made many moves into the Linux space as of late, and have acknowledged that they will be phasing out IRIX (SGI's variant of UNIX) over time. Linux is great for a lot of things, but one place it still lags behind is in application support. This is changing, but it's an uphill battle, and one I am not sure SGI has the strength to fight.

Rick Belluzzo, the successor to long-time CEO Ed McCracken who was with SGI since 1984, threw the company into the NT ring with the Visual Workstation, which caused a loud, painful thud heard around the world. While on paper the system was a design marvel, in reality nothing worked as advertised. Not only that, SGI engineers had to slog through the spaghetti of code that is Windows NT, and create a version made custom for the unit. It never fully worked, and problems plagued the computers from the start. Belluzzo is now an executive at Microsoft, where I am quite sure he can wreak no more damage than anyone else at the Redmond company.

Now the reigns are in the hands of Bob Bishop, one of SGI's largest shareholders. Version 2.0 of the renamed Visual PC have been met with lukewarm receptions, and all but SGI's largest clients like ILM (www.ilm.com), Digital Domain (www.digitaldomain.com), and Pixar (www.pixar.com) have pretty much jumped off the SGI bandwagon. While SGI does own Alias|Wavefront (www.aliaswavefront.com), and Maya is the high-end standard for 3D, the company has been bleeding red ink for the last few years, and it looks like nothing can save them, right?

SGI actually did get it right once, but they failed to follow through. The failure was the Visual Workstation. But the vision part, they got right. SGI's laudable goal was to create a true media super computer, along with video I/O, a fast bus, and fast graphics subsystem. The specs on these NT units were enough to make most 3D artists drool, and many people pre-ordered the units on SGI's claims alone. Although they failed miserably, the world still needs a workstation built specifically for the CG and 3D artists of the world, and maybe the third time's the charm for SGI to make good on this promise.

The first Visual Workstations were too proprietary and not "stock" enough. The second time around, they were little more than clones with SGI skins. This time, the company should focus on what makes the ZX-10 workstations they bought from Intergraph so good, and extend upon that to create a true workstation for the CG power user.

The unit, which I have called The Alchemist, would be a true content creation powerhouse. There would only be one model to streamline manufacturing, but would come with many built-to-order options. There would be no low-end unit because they're low margin, and low-end machines have never been SGI's core competency anyway. The units would all be dual capable, have processors at least in the 1.5GHz range, and come with support for three displays right out of the box—two computer monitors for dual screen work and an NTSC monitor because, of course, the unit would have a fully functional, FireWire based video I/O system. The graphics could be Nvidia (www.nvidia.com) based, or buyers would have the option of something a little more powerful, like the 3Dlabs Wildcat (www.3dlabs.com). The unit would come with a CD/DVD-ROM and a SuperDrive, making the unit very media friendly. The box would have a Zip drive instead of a floppy, and a solid sound card; none of this Creative SoundBlaster (www.creative.com) stuff. The hard drives would be striped into a RAID; about 100GB should do it.

This would make for a more expensive unit than a Dell (www.dell.com) or a Gateway (www.gateway.com), but if SGI could deliver a unit like this at a decent price point while still having reasonable profit, CG and 3D artists would flock to something so customized to their needs. SGI may be on its last leg, but it still has a card or two to play. I just hope they don't screw it up, and I can't see into the future right now because my crystal ball is in the shop.

[doodie]

imagine if SGI became simply a graphics card manufacturer in competition with Nvidia and ATI... I reckon, with the right guy running the show, and making cards to suit everyone's needs (and price points), they could take the crown.

Imagine the Onyx graphics card, or the Octane card.

whew... cool.

[altered]

If SGI do a bankrupt what discreet going to do with there flame,inferno and all that stuff who work only on SGI beast ?

[luceric]

I have an SGI320, it is an extremely stable machine, in fact I've never seen an opengl implementation so stable.
The problem is that the hardware is proprietary, no standard BIOS, so you have to use NT or 2000 and can't boot DOS. In fact, the boot drive is NTFS. The deal breaker is that video card is not upgradable, it's a chip on the mother board. But the performance was excellent, especially with image I/O and it used the computer's main memory, so there was unlimited texture RAM.

Today, you can't upgrade to XP because it's not supported.

SGI got screwed partly because they built the machine around Windows 2000, it had FireWire and USB ports built in. But, Windows 2000 was delayed and they had to back patch everything into NT 4.0, which was a huge problem because it did not support plug and play the same way, and did not support USB (I think) and firewire. This is what delayed the units.
This was also a problem for many laptop manufacturer which had to ship Win98 with their high-end machines.

When Win2k finally came out, it was found that the some pieces like the firewire port wasn't actually working, for whathever hardware or software reason, and SGI had to ship free FireWire cards to everyone. The USB worked, but it is non-standard, since the machine has no standard bios. Not a problem, however, MS ships the SGI HAL with windows 2000.

[update:]Another major flaw : the SGI VPC had a 3v PCI bus instead of the standard 5v. This prevented Avid and other companies from porting their apps on the SGI machine, basically runing SGI's PC hopes of becoming the pro-video PC they wanted to be. SGI needed this position in the market to keep the price higher. They lost everything to Integraph.

I live montreal, and frankly I have no idea where I could have bought an SGI NT machine, this would be the biggest problem for SGI being successful in that volume market -- distribution.
Cool machines though, pretty much the same coolness as a mac, with the big LCD display and all.

It's rock solid. Which is a good thing, because if it breaks there is no way to repair it -- it doesn't even use standard RAM chips.

[swami3d]

You are the FIRST person I have EVER heard to describe a 320 as rock stable, but glad it is for you. I would never question a machine owner's ability to gauge their own box.

I CAN tell you that for 99.5% of all the production people I met that torture tested these units, let's just say they couldn't say what they wanted to..... because there were preschool toys present. :)

ct

[amygdalae]

Originally posted by doodie
imagine if SGI became simply a graphics card manufacturer in competition with Nvidia and ATI... I reckon, with the right guy running the show, and making cards to suit everyone's needs (and price points), they could take the crown.

Imagine the Onyx graphics card, or the Octane card.

whew... cool.

They would just be another 3DLabs - non-standard cards that work great for OpenGL with supreme stability - but are outpaced by cheaper mass market products from ATI & nVidia. There is no way on earth SGI could compete with ATI & nvidia in the gaming, or even workstation market at this point.

[Thalaxis]

SGI is working on a new ultra-high-end graphics card. It's the
next generation version of their Inifinite Reality famliy, I think...
and it uses ATI GPU's.

[Sashelas]

NEC put them in their grave. SGI just can't compete hardware-wise with the 47 billion a year semiconductor giant. It is kind of a 'death of innovation' in the U.S. story. SGI tried doing things like contributing large amounts of machines to internet companies which is foolish since they are a hardware company. They should have stayed focused on their core business.

[Thalaxis]

NEC put them in their grave. SGI just can't compete hardware-wise with the 47 billion a year semiconductor giant. It is kind of a 'death of innovation' in the U.S. story. SGI tried doing things like contributing large amounts of machines to internet companies which is foolish since they are a hardware company. They should have stayed focused on their core business.
Actually, Intel and Microsoft, followed by Intel + Linux, are what doomed SGI. They used to have a booming workstation business... but when Intel processors started matching and later exceeding the performance of MIPS processors at 1/10th the price, SGI's downfall became inevitable... and they obviously realized it and tried to adjust, but unwisely dumped their MIPS projects too abruptly. They would have done much better to follow HP's model of keeping their house RISC alive until Itanium was ready to roll.

[kid tripod]

^^ Agreed, but that said it's simply the desktop 3D capabilities we now have combined with the unix-a-like ness of Linux that has sealed the deal.

SGI needed to stay 64-bit in the shorter term, and try and force home the advantages of that approach over commodity stuff before it turns 64-bit.

Personally I think Apple's G5s are now the threat to what remains of SGIs desktop business. It's a shame because their supercomputers are really really nice. They had some of the best parallel systems people around.

[Thalaxis]

^^ Agreed, but that said it's simply the desktop 3D capabilities we now have combined with the unix-a-like ness of Linux that has sealed the deal.

Actually, I think that SGI's decision to cancel their high-end MIPS projects is what sealed the deal. If they hadn't done that, they'd have had at least a chance.


SGI needed to stay 64-bit in the shorter term, and try and force home the advantages of that approach over commodity stuff before it turns 64-bit.

Exactly -- that's why they should have kept their high-end MIPS projects alive until Madison shipped, rather than cancelling them on the promise of Merced. It could have allowed them to transition more smoothly, rather than suffer from a 3-year period in which their performance was so abysmally outclassed that even Sun was a solid competitor.


Personally I think Apple's G5s are now the threat to what remains of SGIs desktop business. It's a shame because their supercomputers are really really nice. They had some of the best parallel systems people around.
What's left of SGI's workstation market is aimed at a market segment where Apple has never been involved, so it's rather unlikely that the G5 is even relevent to their workstation market. SGI's market for video editing is already pretty much gone in favor of Linux/x86 solutions, and their higher end market is getting replaced by Linux/Itanium solutions, which SGI plans to start selling as workstation products.

[Sashelas]

Good point, but your average consumer doesn't buy SGI's. Governments and high performance computing tech companies do. Memory technology is a different story. I think SGI's fall is largely a historical issue regarding foolish acquisitions with a fickle annual revenue.

In Nov. of 2001, NEC bought 60% of SGI Japan for $95 million. At that time, NEC was doing poorly in the top 100. The very next year NEC opened the 35860.00 Gflop Earth sim topping the list, and trumping all U.S. based labs. It was a good move for the advances in shared memory technology and other things SGI brought to the table.

I'm no expert in reverse engineering, but SGI had some pretty incredible hardware up to that point and was supplying some cool hardware to Los Alamos and other HPRC's.

Also in 2002, Los Alamos switched out to HP.

SGI Japan currently holds the contract for the Japan Atomic Energy Research Institute. Large scale simulations and real time graphics were SGI's mainstay. Their most valuable jewel (research and development) was picked up and improved upon.

Sure NEC was a contender since 1996 when NSF gave them a contract that Cray had their eye on, but the team up, dump and crush strategy of international business is incredibly effective when you want to take the market. If you use a component from another company, they can deny supply or monkeywrench. NEC shorted SGI in MIPS chips in 1996.

SGI picks up Cray, and sees shortages of graphics chips from Toshiba weakening their revenue.

A year later, SGI's are in Russian nuclear bomb design facilities. SGI needed to get their revenue somewhere I guess :)

NSF and DoE should be a little more responsible with U.S. tax dollar turned contract dollar in my opinion since its much cheaper to improve a design then create it initially.

Hitachi and NEC move together to take whatever market is out there. Smart. Smart. Smart. With a combined net sales of over $110 billion annually, they are a pretty brutal rival when it comes to competition. Also, since NEC owns SGI Japan, and SGI owns Cray ($576 million), whatever new computing hardware developed for DoE will probably be trumped by NEC in short time.

If Cray was owned by GE or another U.S. based heavy hitter things might be different. SGI is a victim of large over priced stupid purchases. Executive mismanagement at its finest.

If those executives had kept Jim Clark around maybe he could have kept the spendaholic attitude at SGI under control, and may have made some more intelligent decisions regarding buying critical chips which SGI designed from a rival corporation that wants the market.

[Thalaxis]

Good point, but your average consumer doesn't buy SGI's. Governments and high performance computing tech companies do.

That's [b]exactly[b] my point.

There was a time when companies like GM bought lots of SGI machines. Now they do the vast majority of their work on x86 systems.


Memory technology is a different story. I think SGI's fall is largely a historical issue regarding foolish acquisitions with a fickle annual revenue.

Don't forget interconnect technology. That was what acquiring Cray really brought to SGI's portfolio, and they're still using it to very good effect in their Altix systems.


I'm no expert in reverse engineering, but SGI had some pretty incredible hardware up to that point and was supplying some cool hardware to Los Alamos and other HPRC's.

They did manage to keep some of that market, as well as some markets like GM, but there came a point where their high end infrastructure couldn't make up for having such a pathetic processor doing the heavy lifting. Spinning off MIPS was also stupid -- MIPS has been quite successful selling processors, currently second only to PPC in the high end embedded space, and gaining.


Sure NEC was a contender since 1996 when NSF gave them a contract that Cray had their eye on, but the team up, dump and crush strategy of international business is incredibly effective when you want to take the market. If you use a component from another company, they can deny supply or monkeywrench. NEC shorted SGI in MIPS chips in 1996.

Yet another reason that spinning off MIPS was a bad move by SGI :)


A year later, SGI's are in Russian nuclear bomb design facilities. SGI needed to get their revenue somewhere I guess :)

At least SOMEONE is buying their stuff :)


NSF and DoE should be a little more responsible with U.S. tax dollar turned contract dollar in my opinion since its much cheaper to improve a design then create it initially.
[quote]

True. Of course, the defense industry is optimized to reward the companies that spend the most money rather than the ones who do the best work. If you get your project done ahead of schedule and under budget, your budget gets cut, and the big companies with huge overhead who put out a mediocre, bug-ridden product that isn't finished yet gain. That's about the only reason that LockMart is as big as it is.

NEC obviously agrees also; they announced that they're no longer planning to develop a custom chip for their next supercomputer installation, opting instead for a commodity 64-bit processor.

[quote]
Hitachi and NEC move together to take whatever market is out there. Smart. Smart. Smart. With a combined net sales of over $110 billion annually, they are a pretty brutal rival when it comes to competition. Also, since NEC owns SGI Japan, and SGI owns Cray ($576 million), whatever new computing hardware developed for DoE will probably be trumped by NEC in short time.

Well, it's POSSIBLE that Cray and/or SGI will manage to stay on top. SGI has a rather impressive new supercomputing system in the works, due in about the same timeframe as Tukwila.

What remains to be seen is how well it will stand up to the stuff from NEC and Hitachi. What will happen with Cray is anybody's guess. They just won a sizeable contract for a supercomputer installation, and apparently the customer is paying them for quite a bit of R&D. It seems that they're hoping to commercialize what they develop for this contract.

And SGI just got another sizeable contract as well, for a big cluster that they plan to evolve into an SSI setup. They also got a few wins in the enterprise market (a few companies bought Altix systems for database applications, which is a new one for SGI). That might help them get back into high gear. Maybe.


If Cray was owned by GE or another U.S. based heavy hitter things might be different. SGI is a victim of large over priced stupid purchases. Executive mismanagement at its finest.

Not to mention extremely slow response to a changing market.


If those executives had kept Jim Clark around maybe he could have kept the spendaholic attitude at SGI under control, and may have made some more intelligent decisions regarding buying critical chips which SGI designed from a rival corporation that wants the market.

Of course, there's also the possibility that SGI would only have dragged MIPS down with them if they'd hung onto it. :)

[1001 JediNights]

Do you guys realise you revived a thread that was a year old?

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.