Hot New Silicon Graphics Workstations 472
Jonathan C. Patschke writes: "SGI have finally unveiled their newest-generation visual workstation, the Silicon Graphics Fuel. Features include a MIPS R14k CPU, Vpro graphics, and a PCI bus (finally)." As you would expect from SGI, it looks good,
and the specs are impressive. I only see IRIX listed, but with the
specs on this thing, it may not be slow :)
One of the key apps is Medical Imaging? (Score:3, Funny)
Why? (Score:2, Insightful)
Cryptnotic
Re:Why? (Score:5, Informative)
I wouldn't mind having one of these, but I wish they would bring back their old logo
Want a really fun machine? Get the Origin 2800 w/ 250 CPU's
--Jon
Re:Why? (Score:2)
Re:Why? (Score:2, Informative)
This is obviously important when you can buy 36-bit and 48-bit RGB photo scanners.
Re:Why? (Score:2, Interesting)
Plus, higher pixel depths will allow us to more accurately simulate the accumlation of light in the frame buffer, leading to much more realistic and "correct" lighting solutions.
dual 1GHz MIPS on the same chip ! (Score:3, Funny)
really I dont know why SGI dont use this chip
RM9000x2 [pmc-sierra.com] its got HYPERTRANSPORT like the AMD chips and the ol SysAD bus and Supports DDR SDRAM
all they have to have is GIMP for IRIX ICC'd and most people would be happy for Bitmap manipulation
lots of render's work under IRIX so thats not a problem
the problem is the back end Farm that now EVERYONE uses Linux for on el'cheapo AMD/Intel box's SGI used to live here and now they got shoved out by Linux
they are doing the right thing extend product range and work on getting Linux on decent hardware so they can sell it to their customers
pity Itanium turned out such a PIG
I just hope SGI are doing their own motherboards (-;
regards
john jones
Re:Why? (Score:3, Redundant)
Don't get me wrong, I'm really impressed with the work Apple is doing, but I'd never consider trading in an Octane for a OSX box, let alone a Fuel.
Re:Why? (Score:2)
Not hardly (Score:5, Informative)
This thing looks to have the same terrifying memory bandwidth as its
big brother, the Octane2. 3.2GBps. On a dedicated port crossbar.
The Mac is STILL struggling along with PC133 SDRAM. And the Mac has a
"Geforce4MX", which is basically a faster GF2MX, not a fourth
generation part. Compare that to the SGI graphics subsystem for a
laugh.
For processor bound tasks, yes, the 7455 G4 will be faster than the
R14k, but for overall system performance, ESPECIALLY when pushing big
models around, you'd be goofy stupid to try and use a Mac if you could
afford one of these babies (to say nothing of the Octane2).
Peace,
(jfb)
Re:Not hardly (Score:2)
>>>>>>>>>
Less than terrifying considering the $200 DDR nForce-based boards have 4.2 GB/sec memory bandwidth.
Re:Not hardly (Score:2)
subsystems and then get back to us with "nForce".
Peace,
(jfb)
Re:Not hardly (Score:2)
Re:Not hardly (Score:2)
Re:Not hardly (Score:2)
big brother, the Octane2. 3.2GBps. On a dedicated port crossbar.
Actually, Octane2 has 1 GB/s RAM. Fuel, Origin/Onyx 300, and Origin/Onyx 3000 have 3.2 GB/s RAM.
This guy says it best: http://slashdot.org/comments.pl?sid=27074&cid=292
Re:Not hardly (Score:3, Informative)
To quote [slashdot.org] someone else...
It seems that the GF4MX should be about as fast as a GF3-Ti500, and that's pretty fast.
Milalwi
Re:Moron (Score:2)
Re:Why? (Score:2)
Might point the right direction (Score:4, Interesting)
Because, at least, not everyone should sell Windows machines, let Mr. Dell do it.
Just hope support for Linux up to some extent.
Re:Might point the right direction (Score:3, Interesting)
If you want to argue that he encouraged it, made it a priority, and so on, I'd be willing to listen.
By the way, kids: Mhz is not the measure of a machine. Floating point benchmarks are not the measure of a machine either. The world is a complicated place, and computers are no escape from that problem.
Re:has anyone mentioned price yet? (Score:2)
I hope SGI can pull this off; I still have a soft spot for Irix, even though most of my computing is MacOS X based nowadays.
D
No Mention of UMA (Score:2)
Re:No Mention of UMA (Score:5, Informative)
Re:No Mention of UMA (Score:2, Interesting)
Re:No Mention of UMA (Score:2)
It's probably worth mentioning that Fuel maxes out at 4 GB of RAM while Octane2 can have 8 GB.
Another key differentiator is that Fuel is exclusively a single-user workstation. Octane2 has a dual-head option (two V12 graphics pipes, each with an option dual-channel display adapter) and can take two mouse/keyboard combos. Two users can use an Octane2 at the same time with full graphics functionality. Can't do that with Fuel.
More like lukewarm (Score:5, Insightful)
I work in the video/film post-production business. We are one of their major clientbases, and these machines will go down well in this niche area. Unfortunately, althoguh SGI get a lot of press for their "movie" image, it's not their money-spinner...
SGI get most of their money from government and research contracts. This machine will not cut the mustard in those areas - it's just too damn slow. Yes the CPU is probably a better performer than its Intel equivalent in MHz, but I just don't believe it'll get anywhere near the SPECfp and SPECint of the Athlon 2000 or Intel 2.2GHz CPUs.
It's a shame. I *really* like SGI machines. I've bought several (I donated one of them to libsdl just so SDL would support SGIs
Simon.
Re:More like lukewarm (Score:4, Informative)
This is for Spec CFP 2000 (i.e. floating point). I picked the SGI Origin 3200, which has a similar processor (although I'm not sure if its identical or not).
Advanced Micro Devic Epox 8KHA+ Motherboard, AMD Athlon (TM) XP 2000+ 1 596 642
Intel Corporation Intel D850MD motherboard (2.2 GHz, Pentium 4 pro 1 766 777
SGI SGI Origin 3200 1X 500MHz R14k 1 436 463
The Spec CINT 2000 numbers look similar, I just didn't feel like cutting and pasting.
So, sure your average P4 or Athlon is faster, but its not as simple as a matter of Megahertz.
My concern about SGI is that these machines have the same graphics V10 and V12 that they've been using for years now. I heard that these were designed as the last hurrah of designers who have since gone on to Nvidia or ATI.
I wonder if SGI has the manpower left to design new, innovative graphics architectures, or will they be just slapping more texture and cranking the clock on old designs.
dave
Re:More like lukewarm (Score:3, Informative)
more texture and cranking the clock on old designs.
I've been told that a speed boost along the lines of a "V14" and "V16" will be available in May, with a totally new gfx line (compatible with existing machines as just a new gfx card) becoming available this fall.
Then, of course, there are neat new SGI gfx offerings such as Onyx InfinitePerformance...
http://www.sgi.com/visualization/onyx/3000/ip/
Re:More like lukewarm (Score:2, Insightful)
Re:More like lukewarm (Score:3, Interesting)
I think you're right. There are published results at the SPEC website [spec.org] for the R14000 at 500 MHz. Here's the bottom line (CPU / SPECint / SPECfp - all rates are base):
Looks like SGI should consider joining Apple in the PowerPC world...that Power4 looks pretty awesome!
299,792,458 m/s...not just a good idea, its the law!
Re:More like lukewarm (Score:3, Informative)
That Power4 also costs like $100,000 for each (4-way CMP) processor module alone, so, gee, it'd better be pretty awesome. The 1 GHz G4+ that powers the current generation of Macs would probably score about the same as the R14k on SPEC, or a bit lower...but we don't know because Apple is too cowardly to submit themselves to legitimate benchmarks when they have a bunch of fools running around believing that a G4 is faster than a P4 or Athlon, and Motorola doesn't bother because they know the G4+ is actually designed for the embedded signal processing market, where SPEC scores are not too relevant. Just because the G4 and Power4 are both "in the PowerPC world" doesn't mean they have similar performance characteristics.
In any case, where the R1x000 really shines is in scalability to very high processor count NUMA configurations (not at issue in this case of course). It'd still be a world-class processor line if SGI hadn't given up 5 years ago by essentially stopping R1x000 development and committing to Itanium instead. They've finally realized their mistake and apparently have some extra tweaks on the way (R16k and R18k), but it's probably too little too late.
Were I SGI at the moment, I'd drop IRIX for Linux, port everything that made IRIX special, and run it all on proprietary P4 or Xeon boards with all the special SGI graphics goodies. Although that was the idea behind their NT line and that didn't do so well, did it...
SGI had some amazing tech back in the day, but having more or less rolled over and died the past few years it might be difficult for them to stay ahead of the commidity hardware crowd. (Re: 48-bit color, if johnc has his way--and he usually does--commidity graphics cards will have 48 or 64-bit internal color soon enough.) But they appear to be finally waking up and making a go at it, so best of luck to them.
Re:More like lukewarm (Score:2)
That $100,000 cost is fairly meaningless, since there is an extreme markup on server hardware, and the chip isn't in mass production. (Also this was a single-CPU system, so I don't think it was a multi-CPU module.) I'd venture to say that it can be mass-produced cheaper than P4, as I'll bet it has a lower gate count. The G5 will essentially be this architecture (though I doubt many G5 boxes will have 128 MB of L3 cache like the IBM box).
Regardless, SGI has no qualms about using high-end components or producing expensive systems. It would most likely be good for them to actually be somewhere near the top in performance... ;-)
The 1 GHz G4+ that powers the current generation of Macs would probably score about the same as the R14k on SPEC, or a bit lower
Please cite some reference to support this (wild in my opinion) claim.
Fine, buy yourself a Mac and generate you're own SPEC scores. No one is stopping you, including Apple. (Anyone know if there are high quality FORTRAN compilers for MacOS X?)
Just because the G4 and Power4 are both "in the PowerPC world" doesn't mean they have similar performance characteristics.
What makes you think that Power4 technology won't make it's way into desktop chips? IBM manufactures desktop PowerPC chips as well, and certainly shows no sign of giving up on PowerPC in general. There have recently been rumors of Apple switching from Motorola to IBM for it's chips...we'll see what happens.
299,792,458 m/s...not just a good idea, its the law!
Re:More like lukewarm (Score:2)
OK, how to GET the SPEC tests?!
AFIAK you need about $3000 membership just to get the entire test suite that they use. Not worthwhile to just post some relevant numbers about someone else's product.
Besides, I doubt that that person would want to buy a Mac as they seem solidly convinced that they don't perform as well as claimed when properly comparing the performance.
Re:More like lukewarm (Score:4, Insightful)
don't perform as well as claimed when properly comparing the performance.
I know very well that my iBook with a 500 MHz G3 doesn't perform the same as a 1 GHz P4. However, raw performance is definately not the most important thing in a computer. What matters is that I can do what I want to do. This requirement encompasses an adequate speed, naturally. However, I'm not willing to sacrifice the ability to what I want to do in the way I want to do it in order to gain MHz bragging rights, getting stuck running Linux or (eek) Windows. Some people seem to have other priorities, tyo each her own.
Re:More like lukewarm (Score:4, Informative)
That $100,000 cost is fairly meaningless, since there is an extreme markup on server hardware, and the chip isn't in mass production...I'd venture to say that it can be mass-produced cheaper than P4, as I'll bet it has a lower gate count.
Yes and no. Sure the HPC market where the Power4 currently plays has huge markups and very low production volumes...but that also means designs which could not possibly be cost effective in the desktop market. A single Power4 multi-chip module contains 4 2-way CMP dies, 256-bit interconnect between each pair of dies, and, oh yeah, a measely 128MB of eDRAM.
Each one of the 4 dies takes up 400mm^2 on a
G5 will essentially be this architecture.
The G5 is an upcoming 32-bit embedded chip made by Motorola (like the G4 and G4+), and does not resemble the (64-bit) Power4's internal architecture in the slightest. Whether this chip will be the basis of the next generation of Macs is of course not yet known.
Please cite some reference to support this (wild in my opinion) claim.
Because Apple does not have the integrity (nor, according to the oft-repeated excuse, the FORTRAN compiler) to submit SPEC runs for a G4-based computer, there are no official SPEC scores for the G4. However, we do have Motorola's *estimated* *SPEC95* scores for the 7450 (a.k.a. G4+) at 733MHz. (Here [motorola.com], second page, on the left.)
They are 32.1/23.9, SPEC95 int/fp. By comparison, a 400MHz R12k (best I could find for SPEC95; it is an old benchmark after all) scores 24.2 [spec.org]/43.5 [spec.org] SPEC95 int/fp; 25% worse on int, and 82% better on fp.
That same 400MHz R12k scores 347/343 on SPEC2k int/fp. (Sorry, but no more links; the scores are all available at www.spec.org) Assuming equivalent SPEC95-to-SPEC2k ratios (a faulty assumption, but then again we're using estimated scores in the first place), we get our 733MHz G4+ scoring 460/188(!!) on SPEC2k int/fp.
For a scaling factor we'll use the Coppermine PIII, since it has SPEC2k scores available for both 733MHz and 1GHz configs. 1GHz is 22%/16% faster than 733MHz at SPEC2k int/fp. (If you repeat my calcs, be sure to use the 1 GHz PIII scores using the same compiler version as the 733MHz scores.) So applying that to our "estimated" SPEC2k scores for 733MHz G4+, we get even-more-estimated SPEC2k scores of 563/219 for a 1GHz G4+.
So, a decent spot (32%) better than the 500MHz R14k at int, and a significant bit (53%) worse at fp. Plus the CPU in the new SGI Graphics Fuel can be up to 600MHz and uses DDR and not SDRAM like the one I got the scores from.
So...hope that helped.
Re: the Power4 SPEC scores(Also this was a single-CPU system, so I don't think it was a multi-CPU module.)
SPEC2k is single-threaded. The score was obtained using a 4-way Power4 "Turbo" module with 3 of the cores "turned off". The rather sneaky thing is this gave the remaining core access to all 128MB L3, which means the SPEC score probably overstates single-threaded performance a bit.
What makes you think that Power4 technology won't make it's way into desktop chips? IBM manufactures desktop PowerPC chips as well, and certainly shows no sign of giving up on PowerPC in general. There have recently been rumors of Apple switching from Motorola to IBM for it's chips...we'll see what happens.
Power4 is simply not a desktop chip design. Even using one of the 4 dies in the MCM as the basis for a desktop CPU is a shakey proposition, since they're too big (again, 400mm^2 on
Of course, it may be quite likely that Apple turns to IBM instead of Motorola for the next generation of Mac CPUs (especially as it looks somewhat likely that Moto will exit the semi business in the coming year). But it will not look anything like a Power4.
Re:More like lukewarm (Score:5, Insightful)
I have to disagree. I work for a government research center who is still running some Indys and Challenge S series machines for some applications. The scientists there are worried less about speed, typically, and more about stability and function.
Speed is good, but increased function is even more important, and above all else it had better NOT crash on day 13 of a 14 day modeling operation.
Re:More like lukewarm (Score:3, Interesting)
also know an SGI system will.
The point I'm making is that uptime of less than a year is hardly news. The fact that you have paid an enormous amount of money for something does *not* guarantee it will be proportionatly better in its service than 'el cheapo' replacement, at least not in this environment.
I used to work for the MOD (the equivalent of the US DARPA). We used to have SGI onyx's for our simulations, and life was good. We also had PC's running linux, and life was equally good. There were a few sun's, but because everything else was just as reliable, but got the job done faster, the Suns were marginalised. I could be less subtly about this, but I don't feel like it
ATB,
Simon.
Re:More like lukewarm (Score:2)
4:31pm up 9E99 days, 59:59, 1E50 users, load average: 99.65, 99.55, 99.95
Do you believe that? I thought so...
Long way from my poor Indy... (Score:2)
OT - does anybody know of a Irix UG near Wichita, KS?
Re:Long way from my poor Indy... (Score:2)
Price/Performance (Score:5, Insightful)
This new system looks great but unless I was trapped by some particular piece of software I still could never cost justify buying one. $12000+? Sorry. Even presuming that the real world performance is significantly ahead of a high end Pentium system (which I doubt) it's still more expensive, especially once you factor in the service contracts. Those will add several thousand a year. Not to mention that a "well equiped" version will cost much more in all likelyhood.
SGI makes great machines but as a business they are in a teeny-tiny little market niche that is being eroded far too quickly by commodity hardware. They manage to keep ahead for the super high end stuff but that never leaves much room to grow. Frankly I'm mildly astonished the company is doing as well as it is.
I'd love to play around with one of these new Fuel systems but I doubt I'll ever have the chance. There just are too few cases where anyone could justify buying one. Sad really...
Re:Price/Performance (Score:2)
Last hurrah for SGI (Score:5, Interesting)
I'm not about to enter the SGI vs. Linux vs. Mac debate; look no further than the company's own stock price. Back in September the stock hit a low of $0.31 per share, though it has made impressive gains in recent months due to potential government contracts.
Even in the great technology spending spree of the late 1990's SGI languished far behind everyone else. The company has lost money each quarter since at least 1999, the company is expected to show a net loss for the fiscal year ending in June, and the June 03 year is expected to be breakeven at best. Currently only four analysts follow the stock; jokes about the usefulness of analysts aside, 3 have it rated a hold and 1 has an outright sell.
How much longer will SGI survive. The technology is great, but can they pay the bills?
Re:Last hurrah for SGI (Score:2)
SGI is very plugged in when it comes to open source and Linux. I wouldn't be totally surprised to see SGI try a last ditch effort as a Linux vendor after selling off hardware divisions, if they can afford to do that.
Re:Last hurrah for SGI (Score:2)
You also made the same fundamental mistake that SGI made at one point with Linux clusters. You said "Think NUMAFlex cards for x86 hardware". Suddenly, you are now talking about spinning custom ASICs that don't sell in large volumes. Your harware is no longer entirely "commodity". This drives the price up more than you would think.
sweet design (Score:5, Interesting)
And of course the speed and power don't hurt...
Looks like they skimped on storage features (Score:2)
Re:Looks like they skimped on storage features (Score:2)
Because one of those three drives is the system drive. You can only do striping or mirroring with two disks, and the OS supports both of those in software through XVM or XLV. There's basically no reason to have internal hardware RAID.
This coupled with the fact that no SGI system has ever had internal hardware RAID. It seems that the customers aren't exactly clamoring for it.
Re:Looks like they skimped on storage features (Score:2)
Note the option they list for a Fiber Channel card - that'll be a popular option. Too bad for them it's not ready yet....
Very true. Most heavily used SGI workstations I've seen tend to have either a Media SCSI RAID or a Ciprico FibreChannel RAID hanging off them. And heck, for uncompressed realtime HD video, you *need* that kind of thruput.
I've already asked my sales rep about the FC card delay. Seems the card is available (same FC card used in other PCI SGIs... such as Origin 300) but offical support is delayed. If you buy the FC card now, it "should work fine" but tech support has been delayed until testing is complete.
What a piece of crap!! (Score:5, Insightful)
OMG! Like, for real? Only 600 Mhz max? What kind of slack ass company makes a computer that slow these days? These people are totally lame! {/SARCASM}
Sorry for the trolling guys, hopefully some of you find it funny. I just thought I'd do my impersonation of 75% of the readers when they evaluate Macintosh specs. Anyway, happy modding!
Re:What a piece of crap!! (Score:2)
Of course, there is the Alpha, still whopping ass after all these years, that no mainstream UNIX vendor uses. Why???
OMG! 800MHz Itanium? Totally lame! (Score:2)
But that doesn't seem to stop Compaq and HP from "betting their prospective companies" on IA64, even if it is a dog and they've sold about twelve of them last year...
MHz isn't everything (as Mac people can attest), but I would still like to see SGI start making PC video cards and their own Linux.
...or at least buy NVidia.
Win a Fuel workstation! (Score:2)
http://www.sgievents.com/developer2002/ [sgievents.com]
Time to dig out some old, fun OpenGL code... and maybe gcc too (http://freeware.sgi.com)
Of Course IRIX Only (Score:5, Informative)
I only see IRIX listed
That's becuase this is their latest MIPS system, not some x86 box. Despite some progress [sgi.com], Linux does not really run on SGI MIPS boxes. And some of us like IRIX just fine, thank you :-)
Re:Of Course IRIX Only (Score:2)
Now, I've used Irix before in my daily job (only for about 6 months, though), and I've known at least 3 others that have used it or still use it for work. The key being that we were all essentially forced to use it. I'm not completely dogging Irix, I guess I just never found a reason to like it. It served its purpose decently and only occasionally gave me real problems, but I was still elated when everyone was moved from the SGI machines to x86 machines running Linux (oddly enough, because of management concerns regarding SGI's future).
Re:Of Course IRIX Only (Score:2)
There is a diference between running on a MIPS CPU and running on a SGI/MIPS system. SGI has different bus architectures, sound, video, ethernet, etc. etc. One of the SGIs I have at home even has a built-in ISDN modem! :)
If I understand what I read (and what I overhear lurking in the Linux/SGI channel on irc.openprojects.net) the tough part of about getting Linux running on an SGI is not the MIPS CPU, but all of the other hardware. That said, I still want to try out Linux on an Indy sometime, just to see it.
Now we have to listen to... (Score:5, Interesting)
The machine is nice, SGI makes a fine product, and with renewed violence on the part of the US military they have some chance of being solvent again in the near future. So relax, enjoy looking at a beautiful product you will never be able to afford, and don't be so jealous.
SGI - Corporate Necrotic Agent (Score:2, Interesting)
Prove the speed to me (Score:4, Interesting)
I have one scientist I support. I told him that the p4 was some hot computing (in more ways than one). He put his app on it. His $5k linux machine (dual p4) outran his dual R10k (might have been 12k, can't remember) but 4x. Some might say "Well, ya...that's such an old box". I'll say that it has to last longer because it cost $60k! Not to mention the memory upgrade prices.
There comes a point with the hardware were it is cheaper to get a programmer to optimize your app for a linux machine, or to buy a compiler that can fake out your 32 bit box into doign 64 bit-ish instructions.
Re:Prove the speed to me (Score:2)
Could you provide me with a URL for this motherboard pleae?
Thanks,
Re:Prove the speed to me (Score:2, Informative)
http://www.supermicro.com/product/motherboards/
Re:Prove the speed to me (Score:2)
Go to SuperMicro [supermicro.com] and look at their Xeon boards.
Re:Prove the speed to me (Score:3, Informative)
SGI is losing popularity (Score:2, Insightful)
Been Said Before; So I'll Say it Again (Score:3, Insightful)
I think it's getting to be evident that the traditional UNIX RISC workstation vendors are having a hard time keeping their CPUs not only on the price/performance curve, but on the performance curve itself.
The MIPS chip is battling uphill, just like the UltraSPARC III against competitive offerings like the 2.2 GHz Northwood P4 and the AMD Athlon XP 2000.
I respect SGI for it's history of graphics expertise and devotion to producing quality hardware, but like many others I have to ask the hard question:
For some people, it probably is worth the extra money. But I think that target market is constantly shrinking.SGI has hemorrhaged some good people, money, and their 3D patent portfolio (to MS recently). They can ill afford to come up with any product less than a perfect bullseye at this stage of the game. I fear this is not it.
It's Badass (Score:5, Interesting)
Since this is a workstation, it's primary usage in the post production industry would be as a modeling or editing station, like the Octane/Octane2. Actually, looking at the specs, this looks like an "Octane lite." Note that in the expansion section they do not mention available XIO slots, so no HD (snowball) cards for this puppy. As for the lack of UMA mentioned by another poster, UMA was only ever available on the O2 and x86 visual workstations. Using the system memory for texture is good for CAD applications, but not so good for the real time manipulation of textures needed by Maya or Discreet's compositing applications. Note that the stock graphics are VPro V10 - pretty badass. Personally, I have a V6 in My Octane2. In short, this is an R14k single proc Octane, with no XIO, not too sure about the backplane, as there do not seem to be any fuel related docs up on techpubs yet. For 1/2 the price of an Octane2, this seems like a pretty good deal to me.
Now, as for the clock cycles. Please. Hasn't the recent AMD vs intel clock cycle mess taught you people anything? Clock Cycles !=speed. I mean really, this is not a box to play quake on. This is a box to design quake on. =)
Finally, on a personal note, I think it's pretty amusing that they have returned to the Crimson color scheme.
Good work all around lads, glad to see that there are still enough good people at sgi to get this kind of box out the door. I think that this box is a good mid range system, right between the O2 and the Octane.
Re:It's Badass (Score:2)
I have a PowerPoint that I don't think I can redistribute, but the info in it isn't secret any more.
Think Origin 300 for Fuel's internals. The CPU is connected to the Bedrock ASIC (memory controller and system crossbow) via a 200 MHz 64-bit FSB (1.6 GB per second). The Bedrock interfaces to the RAM over a 200 MHz DDR bus (3.2 GB per second) and to the XBridge, which is another crossbow and protocol translator.
XBridge interfaces to the graphics board, which has a new type of interface that's sort of like a cross between XIO and PCI. It's proprietary, and it's 1.6 GB per sec.
XBridge also interfaces to the two external PCI busses-- one 64 bit/33 MHz with two slots, and one 64 bit/66 MHz with two slots. Also, XBridge connects the internal stuff like the serial ports, Ethernet, internal SCSI bus, and so on.
This is exactly the same internal architecture used in the Origin/Onyx 300 and Origin/Onyx 3000.
Re:It's Badass (Score:2)
The backplane, etc is much faster. Where Octane/Octane2 is based on Origin 2000 tech, Fuel is based on Origin 3000. Fuel RAM is 3.2 GB/sec (Octane is 1.0 GB/sec). Fuel CPU interface is twice as fast. Fuel crossbar switch latencies are about half as long.
Octane is a fully loaded 18-wheeler semi traveling down the highway. Fuel is the same semi, with a smaller trailer, and some aftermarket racing tweaks.
Fuel uses a single, flat board with perpendicular expansion cards... somewhat PC (or Sun Blade 1000) like. There is one XIO interface and that's used for graphics, but the physical connector is not traditional XIO.
Re:It's Badass (Score:4, Insightful)
Well I already knew that before the mess but recent wars between intel and amd has tought me something which most of the people don't seem to get: with enough competition between two strong componies, a product line can evolve to unimaginable heights. The x86 line is fast, so fast that they make everything else seem ridiculously slow or ridiculously expensive or both. The x86s were not designed to be one size fits all, but it turned out they came to be just that.
One can buy a dedicated super-computer for 1000X the price and 100X the power, or a computer 3X the price 2X the power but noone in their right minds should buy a computer 2X the power with a 10X higher cost. Instead one would buy two x86s and match the power or buy five of them and do some weird stuff!
Price/Performance doesn't get in the way if you cannot get the performace you want no matter the price on an alternative platform. Older SGIs were expensive too, but they are one of the few computers that could cut it. You couldn't just buy a hundered 486s and expect same performace. This just doesn't happen anymore except for supercomputers. Current x86s are very fast that there is no offering in the Or perhaps this post is offtopic, as benchmarks quoted here show that fastest single x86s are faster than R14 already.
Re:It's Badass (Score:2, Informative)
Yes it is: http://www.sgi.com/fun/freeware/games.html
it'll smoke any x86 1+ghz CPU even at 600mhz (Score:2, Informative)
Re:it'll smoke any x86 1+ghz CPU even at 600mhz (Score:2)
Look again. I think you'll find quite a few W2K boxes both on the render farms and desktops in the movie biz.
Waddaya mean, Finally PCI? (Score:2)
PCI has been available for Octanes and O2s for quite some time - externally. And they've had a better bus for everything integrated for quite some time too. All you get, in terms of PCI, with this system is an internal card cage.
Too slow (Score:2)
Apparently the above processor is becomming popular for areas other than networking, its intended market.
-Aaron
Folks are still forgetting some major things... (Score:5, Informative)
Folks run IRIX for HD video editing, effects compositing, and 3D modeling for a reason -- it works and it doesn't have the "crap out" effect when working under a huge load. Sure the CPUs in an SGI aren't extremely powerful, but that doesn't matter much -- it's the crossbar switch architecture (Octane/Octane2 is based on Origin 2000, Fuel is based on Origin 3000) and wide busses that make the difference. Batch jobs and long haul rendering is all done on a farm of cheap PC's anyway (unless you're ILM, which owns six Origin 2000s, each with 128 CPUs).
Secondly, SGI is coming up with some way cool graphics offerings. In my opinion, the new Onyx InfinitePerformance graphics is bigger news than the new workstation:http://www.sgi.com/visualization/onyx
SGI screwed up big time in the past, but they're working on fixing the situation. They can't do everything at once, but they're working as hard as they can. They're a pretty wide spread company. Hell, they even own Alias-Wavefront (ever heard of Maya?). They're doing some other cool things, too. Their developer program is now free to commercial developers, but hobbyists with a real project are invited as well in a case-by-case basis. They're even giving away a Fuel workstation at the SGI Global Developer Conference next month. And it's not just a drawing, either. The winner of the machine will be a hobbyist with an attendee-voted best project. Very, very cool stuff.
http://www.sgi.com/developers [sgi.com]
http://www.sgievents.com/developer2002/ [sgievents.com]
Re:Folks are still forgetting some major things... (Score:2)
I still think Irix is the most usable of any X-Windows environment, but it's beginning to show its age. Have they considered tweaking it a bit so it's not so Motif-like?
I still use an Indigo2 R10000 at home. It sure was easy to buy at $400-odd.
D
Re:Folks are still forgetting some major things... (Score:2)
It's my impression that the Althlons and P4s are crossbar switch systems. It may not have as many routes - DECpaq called them "D" chips or something like that. Of course, the number and type of such crossbar switches do have an effect on the ability to move data swiftly.
Am I wrong?
No audio, huh? (Score:3, Interesting)
Digital Audio Through USB ports
Now, I'm not sure how many of you are familiar with this, but SGI workstations are known for their great audio capabilities. Even the humble O2 [sgi.com] has 8-channel 24-bit ADAT optical audio I/O; that's quite something! It seems SGI has decided that this level of audio support is no longer desired, though... Too bad. I'm not sure if USB can be pushed to support this; at 48kHz sampling rate, 8 channels of 24-bit audio requires a minimum of 9 Mbps of bandwidth, which is less than the 12 Mbps theoretical maximum. *Shrug*. Of course, there's PCI slots, but having it integrated was very convenient. And cool, too.
Well (Score:2)
Other readers claim that this machine is "too slow" compared to current technical capabilities. This may be due to the fact that in the 80's, SGI machines were much faster than commodity systems, percentagewise, than they are now. I believe, however, that this SGI machine is just right for science, government, and media, just as SGI claim. My belief is that such organizations have a complete computing environment, so to speak, in the form of a network. Although the enthusiast, such as many Slashdot readers, likes to install fancy shmancy computers on a network for the "power trip", I believe that in a serious workplace environment, every machine on a network has a defined purpose. This SGI workstation fits perfectly in such an environment, where most tasks take place on back-end server machines, leaving the workstation free to process the user's application and display high quality graphics. Besides that, don't forget that SGI's systems offer services and reliability not found on your typical Dell running Windows. Speed isn't the only reason for buying an SGI machine.
Well; that's a deep subject.
SGI beginning to be able to innovate again! (Score:5, Interesting)
For those doing the MHZ thing while bitching about the price, forget it. This is a visual workstation. For those doing modeling, imaging, MCAD, and other graphical tasks, Fuel is hard to beat. There are things that even older IRIX machines do easily that give todays PC the fits. I use them all the time and they are worth what you pay for stability, long life, and capability.
Think of it this way also: You will now be able to get re-maunfactured Octane machines, with very good GFX systems for a lot cheaper in the next coming months. Given the very long life of these machines, that can only be a good thing.
These attributes are what holistic design gives you. Sure the price is higher, but you do get exactly what you pay for... For an example, look at Apple. Say what you want, but they are doing very well while copying what SGI has always done for years. Slowly the 'market' (read: masses) are beginning to figure out that this approach has long term value.
Basically you almost never throw an SGI machine away. When used for one of the specialized tasks they are built for, they continue to be useful long after they should be.
A little off topic, but look at Apple machines and realize that they will be good for making DVDs a long time from now. 5 years from now an older G4 with the DVD drive will still have nice value because it gets the DVD tasks done right. This is how SGI machines have almost always been.
So pay more now, but if the purchase actually reflects the strengths of the machine, you pay a hell of a lot less later.
There is more coming this year I'll bet, it should be an interesting one for SGI!
sgi (Score:2)
unparrallellogram (Score:2, Funny)
("the unparalleled VProTM 3D graphics system for IRIX®")
But Fuel is not the point!!! (Score:5, Informative)
The actual announcement reffers to the so-called Visual Area Networking [sgi.com] - a concept that, basically, boils down to distributed visualising and data processing over a network.
With VAN, a user can interact with an InfinitePerformance supercomputer (usually an Onyx 3000 with several hundred processors), let the big iron do the data processing, and receive the resulting images over a network to a thin client. That "thin" client may be a Fuel workstation, a PDA, some device used by US troops to get realtime maps of the enemy positions, whatever.
The point is, many people, working from many different locations, can work together using their thin clients, but manipulating data on the same supercomputer. I've seen some impressive demos, where two people were immersed into the same VR environment, and were manipulating objects on the same scene, at the same time, over the network. Given the fact that the scene was not just a pure graphical computer-games scene, but an actual simulation with real physical laws and everything, that was pretty damn cool.
I tried to submit the actual story, but it was rejected. Instead, Slashdot caught this ridiculous story about "yet another workstation from SGI". Come on people, get real...
Re:Look Great (Score:2)
Re:Look Great (Score:2)
Re:Look Great (Score:2)
I guess that was what you meant though by saying that a G4 renderfarm would suck.
Cryptnotic
Re:Look Great (Score:2)
$ whois marathoncomputing.com
Whois Server Version 1.3
Domain names in the
with many different competing registrars. Go to http://www.internic.net
for detailed information.
No match for "MARATHONCOMPUTING.COM".
>>> Last update of whois database: Tue, 29 Jan 2002 05:21:04 EST
The Registry database contains ONLY
Registrars.
Re:Look Great (Score:5, Informative)
Re:Look Great (Score:2, Interesting)
Apple's claims are nothing but ridiculous, since they are doing pseudo-FP with their velocity engine. These are 8-bit FP Ops that apple uses for their MFLOPS. SGI is using DP-64bit FP Ops for their MFLOPS rating, so theoretically you should divide appple's number by 8 to get the same FP numbers. So apple's G4 is more like a 0.9GFLOP machine. Theoretically then, the R18K is 3x the speed of a G4 at 1GHz at a slower clock speed. So much for Job's Mhz myth! That is why Apple's claims are nothing but a source of good laugh's when they label their systems as "supers", ooohh look 8-bit FP!
Also a Vpro V12 has a) more color depth per pixel, b) siginificantly larger texture memory, c) Most of the OGL pipeline in HW, d) Does geometry processing on chip... and on and on....
Re:they're doomed (Score:2, Interesting)
1)ILM and Linux? No, I don't think so. ILM uses SGIs, which means IRIX. They've got a shitload of them, and probably MILLIONS of lines of proprietary code, all written for SGI machines running IRIX. And, unless I'm mistaken, ILM has a deal with SGI where ILM gets SGI's hardware for dirt cheap, in return for being a testbed/advertisement for SGI.
2)500 mhz may not sound like much, but remember: it's a 64-bit CPU. All you'll see in a PC or a Mac is a 32-bit CPU. Yes, yes, I know, more bits != better, but neither does more mhz. Besides, SGIs have an incredible amount of memory bandwidth, due in part to their wide data bus.
Re:they're doomed (Score:2)
Doomed? Maybe reborn. (Score:4, Insightful)
There has been great discussion within the "SGI camp" about SGI's abandonment of MIPS and adoptions of x86; many people being disheartened by this. With the new release of this machine, I think it will make many people take a second glance at SGI before choosing an x86-based Linux farm.
Why do people choose SGI? Because with SGI you get a workstation that was designed for Unix; a real Unix workstation. It's an all in one package-- hardware, software, support. It's not some Linux-based kludge.
Look at Apple -- they are nearly identically copying the SGI business model with the release of OSX: an all-encompassing unix workstation solution targetted towards content creation.
While I only own a few less powerful R4400s and R4600s, I believe the R10000 based SGI machines (Purple Indigo2s) are 64 bit... and those were released 8 or 10 years ago-- making moot of your last point. Plus, anyone with any hardware experience outside the x86 realm will note that you are falling into the 'megahertz myth.' Alphas are great and all, but they are being phased out, even though megahertz-per-megahertz they are probably 2x-3x faster than x86 processors.
Welcome back into the ring, SGI
Re:Why not get a Mac? (Score:2)
Re:977 BTU/hour (Score:2)
Wow (Score:2)
Re:I'll take a new Mac, thanks (Score:2)
Of course, there is GNOME (http://freeware.sgi.com).
And IRIX *apps* don't look too shabby...
http://www.ifx.com/pages/piranha/screenshot/dx2
http://www.electronicfarm.com/mule/screen_image
Re:InfiniteReality3 vs. InfinitePerformance ? (Score:2)
InfinitePerformance is a new option for Onyx-class graphics. Its major focuses are faster geometry and lower price... at the cost of reduced texture and AA features.
Both are scalable and can come in a variety of configurations (ie, multipipe DPLEX IR vs multipipe IP). Each has a unique target audience. IR (and future versions of IR) are for folks needing extreme quality and HUGE texture sets. IP (and future versions of IP) are for folks looking for a lower cost option and not needing all of the bells and whistles of IR... but still wanting something way cooler and way more expandable/scalable than desktop 3D.
Re:InfiniteReality3 vs. InfinitePerformance ? (Score:2)
...
IP (and future versions of IP) are for folks looking for a lower cost option and not needing all of the bells and whistles of IR... but still wanting something way cooler and way more expandable/scalable than desktop 3D.
Put another way, IP is more for mechanical CAD guys designing precision-shaped parts (who don't care about the colors/textures too much), and IR is for people in the film or flight simulation type crowds (and other markets that I'm conveniently forgetting.)
Re:Now... (Score:2)
Perhaps best of all, since it's the remanufactured special, it still has the old SGI logo on it instead of the new abomination
The Apple Cinema Display is better, but the price difference is rather stunning, especially considering that SGI is the company noted for excessive prices
I was amazed to see that the 1600SW replacement is just 1280x1024 and is therefore just like any other monitor. I'm surprised they gave up the extra resolution, which they should have known was what made their unit special.
Bizarre.
D
Re:SGIs Are Rock Solid (Score:2)
Well I guess I'd better counter this anecdotal evidence with some of my own. I have used Indigos, Indigo2s and O2s up until about 2 years ago and then switched to W2K. In my experience the W2K machines crash less often - though you might need to clarify exactly what 'crash' means.
Oh yeah, if you roll your own software for a Pentium for the visual effects business and know how to use SSE2 you'll leave the SGIs standing at the starting line.