Slashdot Mirror
New SGI Altix 3000
dlloyd writes "SGI has just publicly announced the Altix 3000 series of computers that can scale from 4 to hundreds of processors, with up to 64 processors per single system image. Processors each come in a C brick that has 4 CPUs. I/O is done though IX and PX bricks (12 PCI slots per brick, IX bricks have a base I/O controler and two ultra 160 disks inside), just like on the Origin 3900 series. Anything more than 8 CPUs (2 C bricks) is connected by R bricks, which route the NumaLink packets between nodes. The NumaLink network is good for an aggregate 6.4 gigabytes/sec to *each* node. That scales as you add more C and R bricks. Basically, you can think of this as SGI's origin 3000 series, except that it runs Linux and has Itanium2 processors. The performance and scalability is like nothing that has ever run Linux and is *far* ahead of the competition. For those of you who wonder why anyone would need a 64 processor Linux machine, many scientific and technical customers prefer running their code on large, single system image machines. Large single system image machines are also less labor intensive to maintain and admin, plus they work much better on code that needs to share memory and pass messages between threads (even myrinet and mpi is glacial compared to the SGI numalink network and running code multithreaded)."
What is keeping SGI afloat? Service contracts on existing machines?
--Pat / zippy@cs.brandeis.edu
I still don't understand why SGI has foregone such a great OS as IRIX. Why go with Linux? Just trendy, or does it really offer advantages for scientific computing?
IX, C and R bricks
The more expensive his LEGO gets...
Tournament Management Online &
Imagine a... Bang!
Karma: The shiznight, mostly because I am the Drizzle.
For a moment I realized I had the chance to make a silly first post remark. Then I hesitated a minute and realized I be more on topic with a mandatory Beowulf remark.
How many keys/sec?
nuff said. Makes a nice webserver one should think ;-)
Seriously - sounds like SGI is trying to stop people moving away from their system. Maybe they'll succeed in the higher end of the market with this.
Is that something you would like to share with the class, Altix?
Scientific computing has allways been SGI's niche. They unfortunately stumbled around the time that Belluzzo took the helm and wasted the entire internet bubble recovering from the mess that caused.
It's great to see that they're finally back and doing some really serious new stuff.
It's a shame though that they won't be running the AMD 64 bit chips, although, I'll be someone is looking into that.
Congrats SGI !
These machines support 512 GB of RAM in one chunk. A Linux cluster might outperform this thing, but you'll need to chunk your data up to fit into the individual nodes' memory. Sometimes this can be a pain in the neck to do, hence the market for something like this.
If tits were wings it'd be flying around.
Looks like the Machine Planet is coming sooner than we realized! We just have to watch out for the bloody Tlielaxians.
Signed,
Pahwindah Dirt
Don't believe me? Why don't you look at dlloyd's posting history. There is none!
From davelloyd.com the story posters website
;)
Work: Field Technical Analyst, SGI
Now don't everyone go submitting their products at once.
"Not knowing when the dawn will come, I open every door." - Emily Dickinson
That's a lie. SGI has not dropped it's MIPS machines or it's IRIX operating system.
Last quarter IRIX 6.5.18 was released. MIPSpro compilers version 7.4 were released introducing C99 compliance and much better compliance with C++ standards. This hardly sounds like SGI is dropping MIPS/IRIX.
I just got my copy of Linux Journal, what, a week ago, and you guys are just now reporting on this? You didn't even steal your "news" from the right source!
--sdem
i always wanted a nice, cozy, warm, ALTIX brick house... i bet i wont even need a fireplace!
I rather like the concept of this...no more trying to pair up older processors when you run across a board someone is getting rid of a few years down the road. I recall getting a couple dual-PII workstations a year or so ago, and finding a pair of matching (and working) processors to put in them was hell...this way I could have just searched Ebay or my parts stash for a single old part.
In SOVIET RUSSIA... erm...NSA AMERICA, the Internet logs onto YOU!
I don't do this for a living, but it seems that $/MIPS is the only benchmark even worth discussing, so shouldn't one be able to put together massive clusters of boxen to do the same thing, only without the SGI price tag?
Correct me, because I'm almost certian I'm wrong.
ceci n'est pas un sig.
Slashvertisement, indeed.
yawn.
you're dead on with your conclusion that this box is anything but new.
not.
SGI did make Intel-based workstations that ran Windows NT/2000 that used all standard ports (ATX motherboard layout, Rambus memory, AGP video, et al)... particularly the 550 workstation. They also made several workstation lines that used proprietary memory and graphics subsystem that also ran Windows NT.
For a time... didn't SGI repackage Intergraph workstations as their own? They also had an Itanium 1 workstation that was nearly identical to HP's and Dell's Itanium 1 workstation... but I don't think many of those were ever made.
Could someone shoot me some info in-regards to when Itaniums will be available in stores?
If I understand correctly.. the new AMD hammer (x86-64) will be marketed towards regular consumers. Is the same not being done with Itaniums?
Too bad the guys over at The Neo Project didn't have a few of these before they closed their key project...
...if only they were affordable and every site Slashdot linked to had one for a webserver.. ;)
Just curious, but did anybody notice an estimated price for various configurations either the 3300 or 3700? I couldn't find any price info on their site.
The futexes are also cursed!
...may be heard less and less as long as processor/system power increases. Our forecast models **MUST** have as much power and memory available as possible. More power = greater resolution in creating forecast models i.e. where it's going to rain. Glad to see SGI is still in the race!
One thing that I have always liked about SGI systems, is that not only do I get a high performance system, but I also get something that looks good design wise. Other companies, such as IBM give me the feeling that I am buying, in equivalent terms an F1 car with the body of a Lada. If I pay top of the line prices, I also like to have something nice to show off.
Jumpstart the tartan drive.
With all those bricks invovled, maybe they should call it the SGI Tetrix.
Trolls lurk everywhere. Mod them down.
Someone tell me why this would not be the "platform of choice" for something like Oracle. At $70,000, it would seem to compete very nicely with Sun. Am I wrong?
One of the reasons we continue to choose high-end SGI systems like the O3k series is because of the way IRIX can do engineering/scientifical computational graphics. No one can compete with IRIX when it comes to things like fluid dynamics calculations. Many other vendors have tried, but none come close to SGI's capabilities. I sincerely hope they have continued their unique software in their Linux distribution.
As Linux gets into high-end systems, it will drive the industry to compete. Allmost certainly, Sun and HP will have to release high-end systems with Linux rather than trying to keep it on low-end only. Otherwise, it will be SGI and IBM only
Now, If a major would start using Linux in an innovative way rather than simply trying to lower their costs. That would help drive real sales.
I prefer the "u" in honour as it seems to be missing these days.
Now that I've RTFM, let me correct my previous comment -- The Altix3000 runs a single Linux image over up to 64 processors and 512 GB RAM. After that, it's NUMA.
It can, however, do high-speed shared memory over all nodes in the cluster, allowing you to store HUGE shared data sets. Here's a link to the info on the memory.
SGI's Altix machines use Itanium 2 CPUs (up to 150 watts per CPU). They have **VERY** advanced cooling subsystems. This is not the MIPS/IRIX Origin series, this is the Itanium2/Linux series.
The three analysts who cover this stock have a hold rating, which in analyst-speak means sell.
Huh? I thought in anyalyst-speak that "hold" meant hold, and that "sell" meant sell. What does "buy" mean, hold? What does "sell" mean, buy?
I'm glad I read your post - all these years I've been misinterpreting these ratings
---------------------------------------------
SERENITY NOW!!!!!!!!!!!!!!!!
Does this thing even run Windows? It would be cool if it didn't.
...that Linux can't scale.
Hell even the bad boys over at Sun can't touch this thing with their UltraSparkies and Slowlaris.
Even though this system isn't designed for OLTP, I'd love to see a TPC-C benchmark done.
My guess is a 64 way would land around 450,000 tpmCs.
Keep it up SGI.
1. Post article to /.
2. ???
3. Save your job!
mstyne: real name, no gimmicks
that none of you assholes have any money for this stuff.
Doesn't Linux have trouble scaling, whats the big deal if the system has 100's of processors, if Linux can't scale to use them all. :-\
I liked SGI, especially when they were proud to be called Silicon Graphics.
SGI's decline is tragic.
This is a well-known practice. Here is what Forbes has to say about it:
--Pat / zippy@cs.brandeis.edu
Well the Altix 3000 is using Intel processors. You seem to be a bit confused on your comparisons. You can replace processors in this system as fast as Intel can spin a new chip.
The other thing is that the Altix 3000 is using standard PC memory (either PC2100 or PC2700) and their 200MHz variant in the near future.
And yeah, the machine costs more than your white-box cluster, but you get what you pay for - the fastest interconnect and global shared memory WITH scalability.
ALl these years you actually have, he is correct. For a variety of reasons, most centering around their own personal and corporate gain in the long term, analysts almost always overrate companies. Here's the usual rundown:
1) When you see a Strong Buy rating, that means they are directly benefitting from this company's performance, and would like you to make them richer please. Sometimes it actually means the company will do well, or some mix of the two, but it's hard to tell.
2) When you see a regular Buy rating, that means the company is pretty neutral with a possibly good outlook maybe. Translate this as a hold leaning towards a buy.
3) When you see a Hold rating, that means dump this stock like the plague.
4) When you see a Sell rating, if you're still holding the stock you're already screwed and it doesn't matter when you do at this point.
11*43+456^2
Stop trolling, and stop being dumb.
They haven't dropped IRIX or MIPS.
They aren't repackaging anyone elses stuff - this is still the most advanced hardware platform in existance, and who do you think it was that made Linux non-laughable on > 32 cpus (check for all the kernel work sgi has done)
Finally, SGI _used_ to sell x86 boxes running NT, and has since quit. And even then, they were nonstandard parts.
Basically, everything in your post was wrong, except maybe the part about SGI getting desperate, and how what has happened to SGI is a shame.
My opinions are my own, and do not necessarily represent those of my employer.
The NewsForge link does mention the price for the 64-processor SGI Altix 3000. And the answer to your question is, if you need to ask you cannot afford it.
The SGI Altix 3300 server is useful as a stand-alone server, a development platform for larger SGI® AltixTM 3700 superclusters, or a node in a cluster using an industry-standard switch.
From the Register
And you've got a website that is resistant to the ./ effect.
even myrinet and mpi is glacial compared to the SGI numalink network and running code multithreaded
Don't mix shitty parallel computation libraries and actual performance. Multithreaded applications without MPI are, of course, faster than anything with MPI, however it says absolutely nothing about:
Contrary to the popular belief, there indeed is no God.
Of course, if the investing public at large learns this, then the analysts will have to readjust their ratings, introducing a new one - "Extra Strong Buy" that simply means buy, while "Strong Buy" will mean hold, "Buy" will mean sell, and so on.
And once the public catches on to that, . . .
---------------------------------------------
SERENITY NOW!!!!!!!!!!!!!!!!
if its not so I can have an OS that runs anywhere. If SGI is so sure of their design, then why not try mass producing it enough so I can afford one.
Not foregone, by any means. Irix is still offered to customers. It's only logical, since Irix is the _only_ Unix variant that can run on a 1024 CPU single-image system. Yes, i am aware of the IBM and HP systems with a similar number of CPUs, but those are not single-image, those are partitioned into many separated smaller systems each one running it's own kernel.
Irix can run a single kernel on 1024 CPUs simultaneously. It's the only one, until now. Linux can do only 64.
I was at Barnes & Noble last night a midnight picking up the new book. It's hard to work today
Des
its no shame, its just a way to go.
:) just like i got a nice very cheap SGI indigo2 now at home :)
i think its a wise choice, SGI can do two things, die or stay alive, this will keep them alive for the moment.
And i love to have some old SGI box with 64 cpu's after they arent in use at the nearest company any more
BSD can do... what?
On how many CPUs can BSD scale? Irix can scale up to 1024. It's the only Unix that goes that far up. Where's BSD now?
SGI like all server companies cannot drop their customer base suddenly and leave then stranded. They will carry on developing IRIX and MIPS systems and softly let them die. Just like HP with Alpha read this (funny) also shame.
Yeah they have done a lot of stuff for Linux and the boxes do contain lots of their own stuff but its a large step down from making their fully own hardware (very novel stuff) and Irix down to using Itanics and doing Linux hacking.
When a company gets down to that level and gives up their uniqueness then they are no longer special (in my hart anyway).
Afterall there is no end of companies about to release multi intanic systems.
Mouse powered Chips, Open source Processors and Lego
The above comment is a cut-and-paste from the R16000 story.
64 I2s is nice and all, but I really want to get my hands on 100 or so AMD ClawHammers. Now that would kick some serious tail. I bet Quake 3 area would run nicely. :-)
SPAM solution made easy: 1 spammer, 5 cords of rope, 5 hourses, and fireworks. Be creative.
You can use the 64-processor version for not only the simulation of, but also for the real-life purpose of melting iron. A double Itanium2 HP ZX6000 is heating up my office like no computer before. When I turned the ZX6000 on, my daughters' self-made art (taped on the wall) started flapping in the warm air. To me it looks like Itanium2 is server room hardware, at least until we get the 130 nm version.
-- Imperial units must die --
The old Russian military secretly experimented with multi-processor machines like this in the late 1980's. This is actually the origin of much of today's "cluster" technology - although the Russians' secret tech was much more advanced stuff. They reportedly were scared off when the machine seemed to develop a kind of "intelligence" of its own. The machine was dismantled and all back-ups destroyed after the following message mysteriously appeared on the system console: "In Soviet Russia, beowulf cluster imagines YOU!"
Most evil is done by good people, and not by accident, but deliberately; motivated by high ideals toward virtuous ends.
Well... our experience with the O2K was a bit different (haven't touched an O3K though so I dunno much about it). With the Origins 2K and NUMA, if you ran large simulations, the distant memory would flat kill your performance so you had to make sure your working sets fit on a single node. Also, even though each processor board in the O2K had 2 CPUs, if you actually ran an even somewhat memory intensive process on each CPU, your app slogged because the memory bandwidth only supported about 1.4X the bandwidth of a single CPU. Another bad habit was that if someone snuck onto the machine when you were running your 64 process job and they fired up something like emacs, it would cause this nasty shifting around of memory that got you into the NUMA state mentioned above. Later, SGI came out with some tools that would minimize this effect though, which helped a lot.
So... in order to support the kind of stuff we did and run it on 64 nodes, we would have had to buy a 128 processor system because we would only run one process per node and all the nodes were two processors.
So, we also looked at the Sun10K. While the cpu to cpu comparison of raw crunch was lower, the memory bandwidth was uniform so the programs behaved predictably and were almost as fast in any case. It also had the benefit of running 1 process (as far as performance per process) wasn't noticably different from running 64 processes (on a 64 processor machine). At the time, the Sun10K met our needs much better.
Now, of course, all those have been shoved out the door and replaced by something even faster.
Initially, it will be available with up to 16 processors, with later support for up to 128 in a single box. Larger single-image systems can be constructed with Quadrics switches. For operating system, it will run Tru64, Linux, or OpenVMS.
EV7 was built for bandwidth - how's this: 8 Rambus channels for 12.8GB/s peak mem bandwidth(6GB/s read or write bandwidth), and 3.2GB/s PCI I/O bandwidth. In addition to this, each chip has 4 interprocessor ports to support glueless multi-processing. Each link supports 6.4GB/s total or 3.2GB/s in a single direction. The chips are connected through these ports in a 2D torus configuration.
With current RIMMs, each EV7 can support up to 4GB of memory, though later this should increase to 16GB. This system is Rambus done right.
This system forms the basis for the 5 Teraflop ASCI Q supercomputer. Let's see an Itanium system match that. Check out HP's press release in a few days.
Ah... forgot about the graphics part... the other was about simple crunching. Yes, SGI has always had top of the line visualization capability. Nothing like playing Doom/Quake in a Cave.
Yes, I may well be unhip and "not with it"; heck it took me about a fortnight to find something which told where the "all your base" came from...
Oops, I misspoke. ASCI Q is 30 Teraflops, not 5.
You guys are all missing the main point!
SGI is the first billion-dollar systems vendor to move their totally high-end million dollar hardware to run Linux, and not just to run Linux poorly, but instead their mega-boxes *require* Linux to performe excellently (unlike, say IBM "Linux/390" mainframes where Linux is not really the native OS supporting all the hardware features and is mostly a curiousity or very expensive Apache server.)
The other vendors, Sun, HP, DEC, IBM have not been nearly as aggressive and are depending on their own UNIXes to remain on their high-end boxes.
SGI is depending on Linux and has tweaked it enough to run huge, 64-way complex NUMA systems. This is a major infrastructure bet on Linux, and (assuming this is a shipping, working product) a huge mark of progress for Linux that it can, today, support this sort of high-end scalable hardware.
We all knew it *could*, in theory, but SGI has invested in making sure that *it does*!
This marks a major shift of SGI to an Intel/Linux pure play. It's not just a bunch of low-end Linux server boxes (which they've done before, and Sun/HP/IBM also do), or boxes that you can run either Linux or some proprietary UNIX. It's a full-scale massive 64-way NUMA SMP server that is optimized to run Linux.
Hats off to SGI, I say.
(I wish they had better business prospects but its hard to do that with a niche sort of product like high-end SMP/NUMA technical computing. We'll see if they can push it into a broader customer base with sufficient application support.)
I wonder how Oracle would do on this sort of puppy?
--LP
This post was incorrect when it showed up in the MIPS R16000 thread and it's even more incorrect in this thread. The fact that it has been modded up is just a sad comment about the meat-wads that wander through slashdot.
Sad but true. I watched EMC kill DG/UX after they bought DG, HP are doing the same with Alpha and Tru64 (nee Digital Unix, nee OSF/1), and SGI are ditching IRIX. All of them are continuing maintenance of their old systems while it's still profitable, but DG/UX is dead, and the others are on their last legs :-(
"The invisible and the non-existent look very much alike." -- Delos B. McKown
hat's off
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
It was from a joke Yakov Smirnoff (comedian) made. The original joke was "In Soviet Russia ... Television watches YOU !"
2 things come to mind:
1 - Are the PX, IX, and R bricks compatible with Lego bricks? (Anyone else thinking of 8-CPU NUMA Mindstorms?)
2 - Did they call them bricks so their marketing guys can say it's `built like a brick shithouse'?
We now return you to your regularly scheduled discussion...
> The performance and scalability is like nothing
....
> that has ever run Linux and is *far* ahead of the > competition.
Pft. I bet IBM Regatta running PPC64 chips still
kicks this thing's ass. Benchmarks
I also work at SGI and although the company is indeed having money problems, it's still a great place to work and most people are dedicated. Also note that the company is 34/100 in the top 100 places to work for. This data comes from the workers, not some random pick. So you can tell that people at SGI like what they're doing. They like their jobs and their company.
.com's are still around? Even blue chips are having problems. In that context, I think SGI is doing pretty well and is still releasing new products! They are still investing in R&D big time.
I'm proud to be working at SGI considering that I came from a very small town with limited IT jobs.
Sure the company is not the best money-wise, but then again, look around you: how many
-- Leeeter than leet
Why is it that everytime something new comes out, everyone has something bad to say about it?
This new server, just like other servers, is a magnificient piece of high technology and will greatly help companies, government and universities.
My hat to SGI, they released an awesome piece of technology and the engineers deserve a high-5 on their hard work.
If it's so bad, then why don't you build your own super computer and then post your anouncement on Slashdot. I'll be waiting in the mean time.
-- Leeeter than leet
If you look at their developer platform for this machine, you see Fortran, C++, and C listed. No Java.
Just a thought for all the Java folk who got so defensive about my comparisons of their language to others. Java is a useful, powerful tool -- but if you want to develop for top-flight parallel hardware, you don't use Java.
All about me
I imagine these beasties must be pretty awesome heating units. Well, at least you are giving a good hard fscking to the Third Law of Thermodynamics. For an electric heater, you get one hell of a lot of free processor cycles :-)
Clickety Click
And since everyone's plugging their favorite SGI competitor's machine, don't forget NEC.
To think that the people that brought you the Cray Supercomputer have been tinkering with the Linux kernel just makes me shudder with delight. Truly things are really getting better all the time! And these are not just minor modifications to the kernel either, making it run on SGI quality systems with scalability over hundreds of processors is incredulous! This is probably the most exciting Slashdot story I've seen in a long time.
Clickety Click
SGI is a great technology company and like other technology companies they don't know how to market their way out of a paper bag. I really hope they can get the word out and sell enough of these systems and keep the doors open.
Just a general question, how much does one of these cost? today. Can anyone give me a website of a reseller that tells you how much for the machine?
--+> Life, is there any?
The Wall Street Journal today said that SGI
was running a proprietary version of Linux.
Does anyone know what that means?
They release new IRIX/MIPS stuff and it's canned as being proprietary/too-slow/too-expensive/beaten-by-my-P4 /beaten-by-a-beowulf/etc.
I'd have thought that all you Linux-boys would be wetting your pants at this latest announcement (which everyone familiar with SGI knew was comming months ago). Instead it looks like most people haven't even read the article yet are eager to shit on this new HPC platform.
The best comparision I can come up with is Average Slashdotter == Import Rice Boy.
Both love their shitty cobbled together kit and lack any understanding of the complexities of such systems. Both will horribly hack up and mod everything they can to the detriment of real performance. And both love to trash high-end stuff they don't understand and which they'll never lay a filthy a finger on.
I Guess that's why I stick to comp.arch & comp.sys.sgi.* for (mostly) intelligent/robust/informative posts.
because it will never be able to scale... all you haters, think your latest ebay purchase that you installed flavorx of linux onto is the hotest damn thing and that no one can beat you because you only paid $199 for the POS system that only streams your mp3 and pr0n files.. give me a break. SGI is focusing on advanced computing, while you have nothing to add other than negativity and a jumping off point for black hats. sheesh...
WTF MAN
And it was wrong when posted in that story too.
Altix uses Itanium 2s. I'm sure those aren't using old manufacturing processes, any more than the MIPS R16000 is.
As for the memory differences: I suppose Sun E15Ks and HP Superdomes are now using off-the-shelf memory? When did this happen?
Well, I have one of these boxes under my desk, namely an SGI 750. It's a dual processor Itanium 1 running 64-bit RedHat. Interestingly, I also managed to install Windows .NET 2003 Release Candidate 1 after a firmware upgrade. Whoopie doo. Might try to install DirectX and I nice game just for the hell of it.
The Register also has a blurb on this. I like the dig at Sun at the end, meow indeed.
I wouldn't worry too much about that...
Since at this point in time the "public" is more worried about abysmal state of the economy and the last thing they want to do is risk their remaining savings on a fragile market.
The second to last thing the "public" would want to do is listen to or other take advice from a bunch of smart ass little pukes who think jumping from one failed dot.bomb to the next qualifies them as having experience with the "market."
One other thing I'd like to see are some Java benchmarks on a machine like that. I'd like to see a WebLogic or JBoss appserver running on a 64-way box like that. I bet it hauls. JRockit (bought by BEA) has a really kick-ass JVM for IA32 machines, but I wonder if it would get confused in a NUMA setup. The Sun JVM is actually really nice on Linux - I wonder how well it would run on this thing.
-nate
"Picture the sun as the origin of two intersecting 6-dimensional ..."
hyperplanes from which we can deduce a certain transformational
sequence which gives us the terminal velocity of a rubber duck
- this post brought to you by the Automated Last Post Generator...