SGIs Linux Future
james outlaw wrote in to send us a story at news.com that talks about SGIs
Linux Plans. We know about their Intel based Linux servers, but according to this article, they plan to
lead with Linux, and only offer NT as an add-on.
When they start shipping those Visual Workstations with Linux default, I'll definitely start desiring one.
One physical RGB->CMYK swatchbook costs about $100, (it can had for less) and they change from year to year. That's just the swatchbook (not software) to convert to CMYK --usable only if you already know the Pantone RGB values. Which you won't since you're using GIMP. Can you color calibrate a monitor to a Heidelberg press or a Lino-Hell drum scanner in GIMP? No. So you can't take a physical rgb->CMYK swatchbook and match chips to your monitor's colors and get anywhere close to what will come off the service bureau/printer's press. Your boss won't like this. Your client won't like this either. Problem.
Even supposing you had authorization from your manager and just had a burning desire to piss your printer off forever (thereby assuring your jobs always get lowest priority) by insisting on using this software without color management where's your integration with a standard pagelayout program --ie, Quark or Adobe's own InDesign? Nowhere. So there's your lesson for the day, now go tell all your friends what you've learned and don't ask such stupid arrogant, questions again --M'Kay?
Now you know I beg Adobe to port their software to Linux and hope SGI can help this eventually come to be.
Visual Workstations (wish I had one) are designed for graphically intensive operations. They are OpenGL enhanced and have high graphical throughput.
The Visual Workstation is a perfect example of a company focusing on their core strength! SGI's core strength is graphics.
If SGI wants Linux on their machines, you can just about bet that they will be the driver (pun intended) behind advanced graphics on Linux - especially their own hardware.
I say more power to 'em.
Codifex Maximus ~ In search of... a shorter sig.
Actually, Sun briefly sold the 386i, which was definitely x86 based. I've actually seen one. We used to have one sitting around the office as a conversation piece.
--
Sun 386i Models
---------------
386i/150
Processor(s): 80386 @ 20MHz, 80387
Speed ratings: 3 MIPS, 0.17 MFLOPS
CPU: 501-1241/1414
Chassis type: tower
Bus: 4 32-bit slots; ISA (3 16-bit, 1 8-bit)
Memory: 8M (documented) physical
Notes: Shared code name "Roadrunner" with the
3.5" floppy. A variant of the 150 had the 250's
external cache(?).
386i/250
Processor(s): 80386 @ 25MHz, 80387
Speed ratings: 5 MIPS, 0.2 MFLOPS
CPU: 501-1324/1413
Chassis type: tower
Bus: 4 32-bit slots; ISA (3 16-bit, 1 8-bit)
Memory: 16M (documented) physical
Cache: 32K
Notes: Shared code name "Roadrunner" with the
3.5" floppy.
486i
Processor(s): 80486
Notes: Code-named "Apache". A very limited quantity of
these were supposedly built and shipped to
customers just before the Intel-based line was
cancelled.
--
If there's a good upgrade path and they keep support for Irix for a long time, users might not resist so much. Remember Sun burned a lot of users, first with the switch to Sparc and then with the switch to Solaris. People got over it.
SGI might have a better time of it than Sun did because people buy SGI for the hardware. People buy Sun for the software.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
You can make the case that Sun has a lot of brand equity in Solaris and should continue it. IBM can afford to sink money into AIX. But SGI's strength really is it's hardware.
I'm not saying they should drop Irix tomorrow, but a gradual shift makes sense. They can position themselves as the natural upgrade path for ISPs and others who outgrow their Intel/Linux boxes.
They've already released their filesystem. All they need to do is gradually transfer other stuff over to Linux and in a few years there won't be any reason to continue new development on Irix.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
It's a cluster of 48 128PE Origins tied together with an 800MB/s GSN (a.k.a. HiPPI) network. The individual Origin boxes are ccNUMA, of course, but the boxes communicate with each other using MPI over the GSN (much like Beowulf clusters do MPI over Ethernet or Myrinet or whatever). There used to be hardware specs on the web, but those disappeared after the China fiasco a few months back. This page gives some details.
--Troy
"My life's work has been to prompt others... and be forgotten." --Cyrano de Bergerac
Bit of a shame in some ways... however, I expect they'll effectively be giving lots of their tech to the open-source community. If SGI are dropping their support for Irix, that wouldn't be very nice to their some of their existing customers, because that means they are forced to port/convert. Some of you might think I'm nuts for suggesting it, but it's true. One of the many reasons Sun have been doing well is that they're quite clearly NOT moving away from SPARC/Solaris.
According to some people I talked to, SGI doing the Intel/NT Visual Workstations hasn't helped with their image with established customers. Didn't help either that they had/are planning to move from their own CPUs to IA-64. (basically, this makes them 'just another' Intel box-shifter in some people's eyes). Incidentally, the current CEO of SGI is an ex-HP (co IA-64 developers with Intel) exec. You'd be surprised what difference image can make. I heard of a Alpha based server Compaq were selling. They then decided to sell the same machine under the Digital brand. Sales increased 10x or something.
--
What makes the Visual Workstations better than a Dell is their very customized video subsystems. I can't quote exact numbers, but the systems have ridiculously video high bandwidth, which makes them usable in video applications that traditional PC's would fall flat in.
Besides the video system, though, I think the rest of the appeal is the industrial design aspect. Cool looking (Non-translucent) cases, and sci-fi looking LCD displays.
I can't speak at how proprietary the hardware is, but if they write open source drivers for it so that they can install Linux, do you really care? Closed hardware doesn't bother me at all, but closed API's do, and if their drivers are GPL (or something close to it), and they don't try to close the thing up like Glide, everything is fine.
And it seems that they really are embracing Linux. Lets not forget their intent to release IRIX's journalling filesystem. This is no small deal: the lack of a journalling FS has been holding Linux back. Quite likely, they will phase out IRIX and slowly contribute to Linux those few high end features which it still lacks.
so that Linux can catch up to the big boys in the FS department. This is no small act, and I do hope that it is executed quickly.
With them releasing Linux boxes, we should atleast get drivers out of them (hopefully free), and hopefully they will support some of the big projects.
Anyone out there use these things for a living?
Did anyone else notice, or has it already been discussed, that in this article Belluzo says they'll be bringing ccNUMA support into Linux? ...
ccNUMA is definitely drool material - even better than Beowulf, and this would suggest that they are planning to have Linux running on machines in the 'Blue Mountain' (~6000 processor?) class
Seems to me that SGI may turn out to be the best thing to happen to Linux since usenet. After all, what have the three biggest (technical) problems with Linux been since it started getting mainstream acceptance?
1. Lack of a solid jfs
2. Lack of high-end graphics
3. Lack of support for *really* big, high-throughput hardware
All three of which SGI looks to be contributing.
I agree with the previous poster - you gotta love these guys.
"Images are incapable of repose." - Bachelard
Sun never made 386 based boxes. The IPC is a sun4c, same as the IPX and the sparc 2. Sun continues to make Solaris for x86 boxen, though its hardware support is, well, specific. Solaris8 won't even support ISA.
I've finally had it: until slashdot gets article moderation, I am not coming back.
How embarrassing ... since I work for Sun. Those suckers aren't even in the field engineer manuals. They must really not want to speak of them again :)
I've finally had it: until slashdot gets article moderation, I am not coming back.
I saw the opening demo. If I remember correctly, they motherboard is SGI with their own memory and bus but using intel CPUs.
The bus is the key. It could move data many many times faster than your typical board. I wish I could remember the specs. I know it was in the Gigabits.
Romans 10:9-10
OK, here is my prediction (despite all the people who think SGI is in league with the devil, microsoft, Saddam Hussein, and Ronald Reagan):
SGI is(are*) the white hats.
That's right, I think that SGI is going to be a major player in making linux kick NT's buttocks. How? I think in 3-5 years they are going to open source most if not all of the IRIX OS. Considering their state right now, wouldn't it make sense that they JUST sell hardware and not worry about the OS part?
As one person I read put it, why buy 1 machine with NT on it for $10,000 when you can buy 2 machines with Linux (LIRIX??) on it for $10,000..... It makes sense for the hardware-oriented vendors (can anyone spell IBM?) to support Linux so there will be more HARDWARE and less SOFTWARE money being spent.
So b***h all you want about offices in Redmond and any other 'atrocities' you want. I think that SGI is going to be the company that comes out with the Linux fix for 256 SMP, an awesome file system, a cohesive GUI (tie GNOME and KDE together with Indigo Magic or something), high-end graphics programs, and enough other stuff to kill NT.
That's right, you heard me: if SGI does what I've described above, NT is dead. Long live Linux/LIRIX.
--
* I don't agree with the philosophy of seeing corporations as a corporate, anthropomorphic entity. It makes them faceless, like Microsoft....
Even if CMYK is a bunch of proprietary, changing ink color values, editing is usually done in RGB. Then, it is converted to CMYK before it prints, because that is the color of the inks.
So why must this conversion be built into the graphics program? I think it would make sense for whatever driver software that comes supplied with a printer to accept RGB color values, and convert them to CMYK at printing time, according to whatever ink the machine is using. Or perhaps Pantone could sell a simple filter program that performs the conversion.
Even if none of this is done, a person right now could take any finished RGB image, and just open it up with photoshop and print it.
Of course, I agree with you that the way things are, photoshop is a better choice in business. A call for standards is definately in order (just like with everything else).
Vidi, Vici, Veni
I just wanted to emphasize that mental ray has been available on Linux x86 and Alpha for at least a couple of years now.
Another major rendering package, Pixar's Photorealistic Renderman Toolkit AKA PRman may already be available for Linux as well. I saw a demo at ACM/SIGGRAPH'98
I just hope that the various 3D modeling and animation packages are ported to Linux as well. The Sidefx Houdini port is a great first step, but I'm hoping that Alias|Wavefront Maya and Softimage|3D are not far behind.
"Yeah well
I realize there are a decent amount of posts about why this strategy is good for Linux, but maybe a closer look reveals why this is so good.
SGI, currently, has two things Linux desperately needs. The first is a journaling file system. I don't think I really need to explain why this is a good thing; all it takes is one bluescreen with NT and you'll understand me completely. SGI's is mature and stable, and has a very good reputation among the workstation community. Nuff said.
The second, IMHO, is even more important. SGI has (again IMHO) the most outstanding implementation of thread-level parallel processing. Almost all the other platforms you care to look at (IBM, older Sequent, Sun) either depend on MPI coding or are designed using close-coupled SMP, which tend to reach their limits quickly. It seems SGI has profited greatly from their acquisition of Cray Research.
SGI has a great thread library which they have mapped to their NUMA implementation, which scales a little better than SMP does (I'll skip the technical explanation here in favor of the point). SGI's extensive knowledge with multiprocessing comes at the perfect time for Linux, which is this very minute undergoing heavy kernel modifications to better facilitate thread level parallelism.
SGI has so much to offer in terms of technical skill that Linux could absorb at this point in time. Make no mistake, this is a perfect opportunity for Linux to milk the expertise from SGI, who needs Linux to survive.
The opinions I post here have nothing to do with my employer.
All systems based on MIPS chips will run Irix. Keep in mind that such systems tend to live far longer than your average PC. I expect to see MIPS/Irix-based boxes in operation (and supported by SGI) well after 2005, probably out to 2010.
However, SGI has publicly stated that they will migrate features from Unicos to Irix, and from Irix to Linux. Eventually, Unicos will fall by the wayside along with the systems it runs on. A few years after that, Irix will be discontinued. But this is many years from now.
What this news sounds like is that since Linux is available at no cost, it will come bundled with the servers (why do people immediately start talking about workstations at this point?). Has NT ever been 'bundled' with the hardware? Is it costing extra anything new?