Got $18m to spare? That’s the market capitalization of one of Silicon Valley’s most glamorous companies this morning, after Silicon Graphics Inc. filed for Chapter 11 bankruptcy. The size of SGI’s debt – at $664m it’s twice the value of its assets – is enough to deter all but the most determined bargain hunter. Apart from a ragbag of trademarks – such as OpenGL – what growth has SGI left to offer?
…Steve want to buy what good of SGI remains before Microsoft did.
Correct me if I’m snorting Drano, but isn’t SGI a hardware outfit?
Microsoft would be making a sizeable leap from an XBox to a full-on graphics workstation if they actually bought SGI.
Of course, such might be sensible, given Apple’s encroachment into Intel-land.
Where are Dvorak and Cringley with some ‘analysis’?
Well, think about: Mac OS X and Linux is actually the more serius enemies of Windows Vista. Mac OS X uses OpenGL strongly. And even Linux uses OpenGL.
If Microsoft will buy SGI it’ll control OpenGL… Think about…
Edited 2006-05-10 01:59
| Well, think about: Mac OS X and Linux is actually the more serius enemies of Windows Vista. Mac OS X uses OpenGL strongly. And even Linux uses OpenGL.
If Microsoft will buy SGI it’ll control OpenGL… Think about…
Incorrect.
The OpenGL standard is maintained and governed by the OpenGL Architecture Review Board (ARB), see http://www.opengl.org/about/arb/overview/
Remember the question of Vista’s support of OpenGL (e.g. OpenGL layered on top of DirectX vs direct hw access)? Although Microsoft may be powerful, the majority of the world’s bigger companies all rely OpenGL in some shape or form and Microsoft would not get away with a shady approach.
I worry about the OpenGL trademark, though. On opengl.org, it does says that it is a trademark of SGI.
I wouldn’t put it past microsoft to buy the trademark just for the sake of causing marketing problems to opengl.
maybe the only thing from that company that might be worth somthing is the the name SGI.
Edited 2006-05-09 22:47
I wrote this elsewhere but it bears repeating. My advice to SGI is the same one I give to Sun Microsystems:
Don’t try to beat either Intel, Linux or Microsoft at their games. You will lose. I suggest instead that you do something that will take the rest of the industry completely by surprise. Invest your remaining resources and passion into the next big thing, the one thing that will solve the nastiest problem in the computer industry today: unreliability. Put all your money in non-algorithmic, signal-based, synchronous software. This software model will revolutionize both the hardware and the software industry and usher in the most dramatic change in computing since the days of Charles Babbage and Lady Ada Lovelace. Don’t say you weren’t warned. ahahaha…
You mean they should get back to what they did when they made some money. You mean they should …
INNOVATE?
What a bald move. But hey that’s the only way for them to survice. “Me, too” just ain’t good enough.
There is an extreme risk for a company to be that ambitious while in that much problems. Sure they may be able to do it but being the only player on the block no one will care or want it, or wait until Apple or Microsoft or Intel releases it. People really don’t want new technology they want proven ones. And the costs of porting software to such an architecture would be high, very high, because it would require people to rethink their processes.
They should have just brought out the R18000 instead of canning it at last minute. It offer double the flops of their existing MIPS processors at the time, it was ready for production when it got canned in late 2002/early 2003. Likewise they canned IR5 and went with ATI gpu’s they dug themselves into a hole with the whole Windows NT fiasco and they been trying to dig themselves out for last 8years.
SGI spun off Cray in 1999
They axed of the MIPS division between 1998 and 2000 and moved to the Itanium
In 2002 they sold 62 million worth of graphics patents to MS (SGI helped build the N64)
Today they sell mostly Linux systems with off the shelf ATI graphics cards.
I have heard they lost many of their best employees also.
I believe the only things they have left of value are the Irix operating system with has begun to fall behind, and a pile of governemt contracts which have kept them afloat over the past few years but are too temporary to justify aquiring the company for.
Either way, I hope who ever acquires the company does so for good reason rather than just preventing their competitors from getting the technology.
I wish it got converted to run on intel/AMd cpu:s.
Irix is one of the most stabile OSes i have ever used. We actually sold a lot of SGI servers a couple of years ago for the graphical market.
We also sold (and do sell today) SUN, macosX, windows and linux as server OSes, but none of them were/is as stabile as Irix was/is.
To bad they hardly releases any updates for irix anymore.
One of the reasons for that stability is that it ran on very specific hardware which had been designed from the bottom up by the same company who wrote the OS. If they lost the hardware control (as they would switching to Intel) they would lose much of the stability.
Also Irix is written from the ground up for MIPS and has been tested for years on MIPS hardware. Porting would be a massive undertaking requiring re-writing large parts and thus introducing plenty of new bugs.
So even if porting Irix to Intel would be feasable, doing so would lose you the main advantage of Irix.
IRIX was not written from the ground up for MIPS. It is a System V-based OS, meaning that the underlying codebase is quite portable.
It’s Sys-V based, but very large chunks of it is completely re-writen in MIPS assembly and with MIPS specfic features. Irix is not an easy port.
The whole point of the article was that it implied that Rick Belluzo was paid by MS directly or indirectly to sabotage UNIX first at HP then at SGI.
It is up to you to decide.
The whole point of the article was that it implied that Rick Belluzo was paid by MS directly or indirectly to sabotage UNIX first at HP then at SGI.
It is up to you to decide.
If the choices are big conspiracy vs the guy being stupid and incompetent I choose the latter over the former.
Hanlon’s Razor : “Never attribute to malice that which can be adequately explained by stupidity.”
“If the choices are big conspiracy vs the guy being stupid and incompetent I choose the latter over the former. ”
If you choose the latter…. I guess he’s found a good place suited to his talents, then, at MSFT. He’ll be surrounde by like-minded dimwits. I guess either way he is in the right place….
I knew Belluzzo when we were both at HP, and I’ve known SGI since NASA bought half the workstation they made the year they started shipping workstations. I also know the history of Cray Research in great detail, having worked with them as a first line customer at NASA, and having a wife who worked for them.
I’ve also had direct first hand dealing with Microsoft and have a fairly good idea of what they’re capable of.
It would be kindest to say that Orlowski is known for making things up and leave it at that.
Belluzzo’s not that kind of guy. He never sabotaged HP/UX and was very much against the push inside the company to move to NT. He failed at SGI because he didn’t fit their culture very well and by ’99 they were already passed the point of being salvagable. Too many mistakes had been made, too much talent had left.
SGI should have never bought CRI. CRI was a dying company and beyond salvation. The current “Cray” is merely a sales company, pushing other people’s products. Cray had made several disasterous mistakes, starting with John Rollwagon firing Seymour, and ending with the T3. The market for supercomputers was erroding and players were dropping left and right.
OpenGL is an open standard, ever wondered what the Open in OpenGL stood for? Now you know… OGL development is pretty much controlled by a steering commitee, not SGI. In fact most of the OpenGL 2.0 extentions were made sans SGI involvement almost altogether. SGI gave away most of its graphics IP long ago, they are an irrelevant player right now. Sad, but that is the situation…
Yes, I know that OpenGL was an open technology, but I thinked it made a strong use of SGI patents. If these patents are irrilevat now all things change. Happy about this. 🙂
P.S. Please, don’t use OGL for OpenGL. OGL is the Open Gaming Licence, the free distribution licence of the d20 System, the roleplaying system of Dungeons & Dragons. 😛
Edited 2006-05-10 19:16
All I I ever dreamed about was IRIX for LINUX… more mature and, in my opinion, better than KDE or GNOME.
IMHO Jb
Imagine if they OpenSourced IRIX? Would people still care? I imagine not now, but “in the day.”
All I I ever dreamed about was IRIX for LINUX… more mature and, in my opinion, better than KDE or GNOME.
FYI IRIX is not a Desktop Environment. IRIX is a UNIX Operating System.
“FYI IRIX is not a Desktop Environment. IRIX is a UNIX Operating System.”
I do realize that, I wish the desktop environment component was ported…
Also Irix is written from the ground up for MIPS and has been tested for years on MIPS hardware. Porting would be a massive undertaking requiring re-writing large parts and thus introducing plenty of new bugs.
When was Irix written from the ground up, let alone “for MIPS” ? Most of the kernel code is SysV (with a touch of BSD for network code). There are the additions SGI made (NUMA, xfs, etc) but there doesn’t seem to be anything “for MIPS” about them (i.e. assembly code, etc).
So, would you care to enlighten me about these two points (“written from the ground up”, and “written for MIPS”)?
[Most of the kernel code is SysV (with a touch of BSD for network code). There are the additions SGI made (NUMA, xfs, etc) but there doesn’t seem to be anything “for MIPS” about them (i.e. assembly code, etc). ]
Actually, you are only partially right. You are right in the fact that it is based off of SysV and BSD, but there are MIPS specific stuff in IRIX.
There are pieces that are written in assembly for performance issues. Furthermore, there are MIPS specific stuff related to addressing and paging. Consequently, any port of IRIX would be quite complicated.
Actually no, just because it implements a SystemV interface, it does not make it automatically portable.
Irix kernel is quite specialised, specially after the introduction of the Irix 6 cellular scheduler. There is tons of code for real time and NUMA scheduling and memory management, also big chunks of the OS itself are written in pure assembly for performance reasons. The OS is even tuned for their xserver….
They tried to do a port to IA64 in the late 90s, but they realised the resources needed to do so where larger than what they could afford. The OS has been pretty much in life support for the past 4 years or so. It is still quite a remarkable system though, things like its memory foot print are amazing when compared to late kernels such as Linux, Solaris, or OSX.
from now on its apples versus pears 😉
I’m so thankful for having more options (-: