Slashdot Mirror
World's Fastest Macintosh Cluster
gabeman-o writes: "The Grupo de Lasers e Plasmas has created the fastest Apple G4 cluster. The cluster runs on 16 Dual PowerPC G4/450, 32 processors, 12GB of RAM, .5TB of space, and Mac OS 9. Apparently, they have utilized the AppleSeed technology developed by UCLA. According to the website, the cluster will be used for simulating plasmas. Not too shabby!"
For desktops, maybe. But clusters have completely different economics. The "training" cost is *zero* after the first (or second, depending how you look at it) machine. And the purchase price is perfectly linear, so it becomes a far more important concern.
> Wouldn't it be cheaper to do this with a bunch of PC's in stead?
> Not as pretty but it's someones tax dollars that pay for these...
Fact of life: hardware is cheap. A $3,000 Mac vs. a $2000 PC is peanuts, even if the total difference is 30-40k for the whole installation.
Re-tooling the development environment, retraining IT and programming personell, porting code to the new platform and the cost to set up and configure the new and cheaper hardware far outweighs the cost of the hardware itself. Then there's the time lost in accomplishing all of the above.
If your admins know Mac, and your coders and QA staff are used to building and testing Mac code with Mac tools (MPW, Codewarrior, Macsbug), you go with Mac. End of story.
SoupIsGood Food
Why does everyone build clusters using normal desktop boxes? Rackmount clusters make so much more sense. Admittedly, I don't think the G4 comes in a rackmount configuration, but surely that makes other options (Sun Netra, Compaq DS10L or noname Intel boxen) more appropriate?
"The invisible and the non-existent look very much alike." -- Delos B. McKown
Take a look at that picture on the right. There are two monitors in the frame. Look at the right monitor. Look closely. If I'm not mistaken, that's OS X.
Default OS X background, Dock on the bottom of the screen. Oh yeah, they're using OS X. In a way, this makes sense. They say they're using Pooch, and Pooch is available on OS X.
Very very cool!
Kindof a valid point, tho it's fairly obvious
that said cluck is just a bitter old VMS user.
K.
-
-- Proud descendant of semi-nomadic cattle-herders.
With Appleseed clusters you don't need some convoluted set of instructions to set it up, the nodes need only TCP/IP and file sharing turned on and you can use POOCH to control them. You can have a mild mannered room full of graphic workstations during the day and turn them into a super computer at night. Who in the holiest of all holy fucks cares if this cluster isn't on the Top500 list? Supercomputing folk will laugh in your face if you mention Beowulf in the same context as Cray you jackasses. By the way for those of you who are complete fucking idiots, they maintained 50 Gflops using AltiVec aware software which can pump out several values per instruction (SIMD in case you're wondering). If you're going to be using a cluster like this there's no reason not to enable every hardware advantage you possibly can on each of the nodes. As for compatibility you can use MacMPI for C/C++ and FORTRAN code or use distributed object coding in Cocoa in OS X. I guess Linux folk don't appriciate things that actually work how the documentation describes.
I'm a loner Dottie, a Rebel.
Wouldn't it be cheaper to do this with a bunch of PC's in stead? Not as pretty but it's someones tax dollars that pay for these...
Well, due to the fact that the PPC7400 has an Altivec unit it is much faster than any comparable Intel processor.
To get the same speed from Intel hardware you need something like 50 dual Pentium
And the price for rackmount computers is almost comparable with the Apple G4 dual 450.
Second, gigabit ethernet is standard on those machines.
If they want to they can replace the 100baseT switch with a gigabit switch.
Sorry, I got ranty on this. To summarize, the parent article makes some good points, the previous replies are too extreme in their criticism. I supply better anecdotal evidence than them. :-)
For CFD the altivec units will be heavily used, it is correct to use them for comparing speed to PIIIs. There may be a case where it matches 50 dual PIIs, but that sounds extreme.
I just bought three lowend Gateway 2U rackmount boxes. They come in at ~$2200 for a single 800MHz PIII which will be slower than the dual 533MHz mac at ~$2500. That certainly meets `almost comparable'.
Neither of these machines are very cheap computers. There definately is a quality problem when you get into very cheap computers. There are applications for them, but if your diagnostic time is worth much, then I believe it pays to buy better hardware. (Plus you don't bleed when you have to open them up and add memory or drives. Once you get above the nasty low end machines they take time to deburr the stampings. My brother runs a metal working factory and their worst injury this year has been an IT guy that stumbled while holding a cheap PC and sliced several fingers to the bone.)
The new Macs do come with gigabit ethernet (although I don't think those old 450s had gigabit) and an OS. That gateway price is a bare machine. I put Linux on `for free', but the 2.4 kernel series (which I needed for iptables) had bugs in the interrupt routing for the chipset and it took me weeks of effort to get it all worked out into a stable configuration (manifested as a AIC-7xxx problem, took a while to find the interrupt controller problem). I knew I should have bought a 4th scratch box.
Gigabit ethernet hubs are still $250/port, but that will come down quickly. I remember when I bought a Powermac and thought it was just plain silly to put a 10/100 adapter in it, only servers could afford to have 100mbit ports. In 12 months all my new hubs where 100mbit.
I've got loads of Apples and loads of PCs. The apple hardware failure rate is less than half our PC hardware failure rate. Depending on the cost of a failure (in terms of ruined work, lost work, diagnosis, and repair) I do lean toward Macs because of their better reliability. (And before we get into a flame war on your reliable PC... I'm sure there are reliable models of PCs, but I need a vendor I can count on to make every model I might buy reliable. I can't look at historical data and say "look, here is a model that was made two years ago that was a good one" because I can't buy that anymore. (And yes, I own a PB5300, I know the counter argument, but Apple fixes it for free whenever it breaks so thats not so bad.))
--
Although Mac OS9 doesn't natively support dual processors, application support for duals can be added via a system extension.
Several Mac OS9 apps are specifically coded to use the second processor... Adobe AfterEffects, Adobe Photoshop, Cinema 4D, and other graphically intensive apps. Users of those applications appreciate all the cycles they can get.
Since the second processor is dedicated to the single application using it at the time, you usually see pretty high effeciencies for a kludge-- 190% speed increases in those specific apps.
Apple has flirted with duals since 1996's 9600 MP, but it wasn't until OSX that they really made sense.
Marc Siry || interactive media professional, motorcycle enthusiast ||
Classic Mac OS can use multiple processors, it's just that the applications have to be specifically written to use them, which the "clustering" software here apparently is.
I guarnette any admin that is worth his weight can keep 16 X86/Sparc/Alpha boxen running as stable as 16 Mac boxen...
Sure, but if it only takes the admin half the time to manages the Macs, that means he has more time to do other work, or that you can double the number of Macs without hiring another admin.
How to solve most of our problems: 1.Lots of nuclear plants. 2.Cure aging.
I doubt that. But even if your premise is correct you'd have to add the time for rewriting all their software for Windows or Linux. Let's not even start thinking about if some of that is commercial software.
...since you can't buy these machines in any other casings that tower or cube. I doubt that it pays off for 64 machines.
How much extra does the case cost? It's just a few pieces of cheap plastic. And the extra features (handles & door) makes it a lot easier to work with than the standard beige box.
I see a lot of posts complaining about the cases, of all things. I don't think this crowd is really upset about the cost. I think it's that they just can't stand functional things being beautiful. And that's very sad.
When I competed in a sceince fair back in March, amoung other awards I won the "Princeton Plasma Physics Award", an award sponsored by the Princeton Plasma Physics Laboratory (PPPL) here in New Jersey, and as a winner I was given a free tour of the facilities. For those of you who don't know, PPPL is "The hottest place in the Universe", lying at the forefront of Nuclear Power and Plasma research...very cool.
.... MACS? The engineer giving me the tour explained that it was in the personal interest of most of the researchers. Yes, there were Sun's and other UNIX boxes scattered on the control room floor, but I would look closely, and sure enough, amounst the three or so monitors at each workstation, one of them was hooked up to a mac. There were g4's and g3's scattered all across the floor. Wack.
So I was going around the facilites, visiting their $125,000,000 tokamac's and torsotrons and all this crazy equipment (very cool science plamsa physics is, too much to elaborate on here), and I get to the control room, from which they run all their Data Acquisition (DaQ) and such to monitor the expirements, and the room is filled with
So yea, Mac's are playing a key role in plasma research, helping achieve effecient fusion, one step at a time.
For another cool plamsa physics project (unrelated to mac's), check out Garrett Young's ISEF project Quasi-Elliptical Torsatron - A Study of Induced Radial Electric Fields and Plasma Turbulence. He is a senior in high school and on the cutting edge of plasma physics research. Quite the talented individual.
my company has several g4 systems an order of magnitude more powerful (320 CPUs) both in house for testing and in customers hands. and they only eat up 9u in a 19" rack.
and as far as ppc architecture clusters go, just about any IBM RS-6000-SP2 system is more powerful, and probably takes up less space too.
The difference between Theory and Practice is greater in Practice than in Theory.
You mean now that some new piece of whizzy kit gets released we've got to speculate what a Beowulf "distributed parallel computing array" of them would run like? Maybe if someone can come up with a cool-sounding name then folks will stop calling them clusters : )
6 CPU Cray T3E= $630,000
I can't even find a quote on a 32 proc T3E much less a 100(128). Nothing realistic price-wise here.
Insanity is the last line of defence for the master diplomat. But you have to lay the groundwork early.
Are you referecing the fact that MacOS 9 doesn't have Dual proc support... do they realize the second cpu (and the cash used to get that extra cpu) is just sitting doing nothing but burning up (the cash) and causing heat (the proc)?
It's like running a dual proc P3 and using Windows 95/98. Atleast go with something that has dual proc support like MacOS X, Windows NT or Linux... if your going to play with 2 procs.
No, this isn't a flame/troll... just wondering what there logic is behind this... maybe you can code an application by hand under a OS that using more than one proc even if the host OS doesn't support it?
"`Ford, you're turning into a penguin. Stop it.'" -THHGTTG
Good point and all... but how exactly do you save money by buy a SECOND CPU that the OS (Mac OS 9) doesn't even see... won't single proc mac's or an OS that could use them be cheaper?
Last time I looked apple didn't have a "buy one get one free" deal on their CPUs
Also we are talking Professinal Admins here, any decent admin should be able to handle a 16 boxen, no matter want the hardware/software is as long as that is the admin's "grove"...
I guarnette any admin that is worth his weight can keep 16 X86/Sparc/Alpha boxen running as stable as 16 Mac boxen...
"`Ford, you're turning into a penguin. Stop it.'" -THHGTTG
a) When G3s came out, they were marketed as mini-supercomputer.
b) When Mac OS X came out, they said 'your old G4 isn't fast enough to run OS X'
Meanwhile, my 486/33 sub-notebook (which cost me exactly $20) is running smoothly on emBSD (32meg tiny OpenBSD derivitive) on a 100 Meg hard drive, and 12 Megs of RAM. Surfing the web over my cablemodem with Links, and laughing at all you Mac fanatics. And don't even get me started on my Psion 5mx (www.psionusa.com)
---=-=-=-=-=-=---
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
---=-=-=-=-=-=---
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
---=-=-=-=-=-=---
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Researchers at the Grupo de Lasers e Plasmas (GoLP) achieved the first milestone of the GoLP simulation program on Extreme Plasma Physics: the installation of the first Macintosh G4 cluster in Europe, called epp (or ep2), which is based on the AppleSeed paradigm developed at UCLA by Viktor Decyk et al. The epp cluster is capable of delivering over 50 GFlops of peak power, and it is based on 16 Dual PowerPC G4/450, 32 processors, 12 Gigabytes of RAM, 0.5 Terabyte of hard disk space, running Mac OS 9, over 100 Mb/s Fast Ethernet, switched by one Asanté Intracore 8000. This is the fastest Macintosh-based cluster in the World. The installation and set up of this cluster took less than 1 day (including moving the machines to the computer room, unpacking the machines, and making all the cables!), and it did not require previous knowledge of networking: a one-page recipe for Mac OS clusters can be found here (AppleSeed website) (Portuguese translation coming soon). This "supercomputer for the rest of us" will be used for the numerical simulation of plasmas, novel plasma particle acceleration schemes using ultra intense lasers, and relativistic shocks in astrophysics. This work supported by the Fundação para a Ciência e Tecnologia. More info available soon (also in Portuguese), as well as science using epp. For more information also contact Luís O.Silva(+351 21 8419 336).
---=-=-=-=-=-=---
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
if that optical mouse works on the glass table :p
I'd love to see a Beowulf of... oh, wait. Never mind.
/Brian
No one ever did, and Mac users are a loyal bunch. The hockey puck was one of the worst mouse designs Apple ever cooked up, just as the current optical mouse is one of the best (I still won't use one, though -- I need two buttons).
/Brian
Geez, give somebody a bridge to live under...
First off, there was a good reason to hype the optical mouse: everyone was laughing about the hockey puck and Apple wanted to show the world they'd not only fixed the problem, but they'd found a more elegant way to do it. And frankly, I hate the Intellimouse anyway -- a mouse does not need to be that big. (And fwiw I think Sun and/or Mouse Systems was doing optical a long time ago -- they just needed special mouse pads at the time.)
Built-in Ethernet first showed up on the Blackbird (5xx) series in 1993 or so. It disappeared from the 5300s and was brought back in the 1400s and has been standard equipment in every Powerbook since. I don't know if they invented it, but they were doing it a long time ago.
The rest of your statement is too silly for a response.
/Brian
And I am a registered user and you're an AC -- who's handwaving?
On the Mac, you don't have to jump through hoops to take over the processor, you just do it. That makes it rather tricky at times to deal with when writing end user code, but it's no problem at all if you need to run the system headless. You run your code with no interference whatsoever from the OS. (That's how much like you get a non-MacOS operating system run on a pre-PCI system, btw. You simply write a "bootloader" that takes over and says "everybody out of the pool", then install your own code over the (unprotected) memory space you just cleaned out.)
You know, calling someone an idiot in a technical dispute isn't really becoming. Why don't you step out from behind the LoserHandle and explain yourself?
/Brian
I have a 6500 minitower myself. I think it's pretty slick for beige and I keep it on my desk...
/Brian
And 'they' can even do that on OS X, of course...
Actually, I don't believe Sherlock does regexps (though it's not half bad as a search engine), but a lot of editing programs on the Mac (especially BBEdit) do. It's not quite the same, but hey...
And as for burning CDs... You might want to look into, say, Toast (3d party, but...) or iTunes 1.1, because I have a Yamaha CRW8824 that might take issue with that...
/Brian
I rather like your statement about old-line Mac users, though we Mac folks are inclined to consider IBM the good guys these days :-)
Those AppleSeeds are actually pretty nice hardware; I've read the technical report on the original and it's actually a better design than a Beowulf. MacOS Classic's cooperative multitasking is actually a plus in this arena as it means that your computation code can completely take over the system and grind away. If you really want to have some fun, though, write a lightweight bootloader that emulates just enough MacOS and MPI to get the job done; you don't need the GUI overhead for a cluster anyway.
I would make two changes to the design, though: I'd use OS X as a front-end system (since you're not doing much computation on the coordinating server, I'm sure) and I'd start calling the clusters Apple *Orchards*...
/Brian
On the Appleseed page, there are graphs in which they usually include the Cray T3E 900...including "single processor performance." I'm sorry, but...huh? When have you ever heard of a T3E with just ONE processor? Or 4 or 8 for that matter?
Okay, I get that they're trying to make it look better by comparing it with something completely out of it's league, to turn people who would just be using a bit of time on some big iron to having their own personal power playground...but lets be more fair, and see some stats comparing 100 Mac clusters to 100 CPU Crays. More realistic.
I usually consider a G4 to be 2x as fast as a PIII or Duron, clock for clock. Details in what you're doing and who coded your app make a huge difference, however. So do things like having enough RAM to never, ever hit the HD.
The G3 is essentially just as fast as the G4 for nonAlitVec ops - although it was actually MORE expensive when you could buy both at comparable speeds.
Tbirds and PIVs are faster than PIIIs, duh. How much? anywhere between nothing and 50%, probably.
AFAIK, the G3/G4 already has a 64bit pipe - and imo that accounts for most of the big differences. Most of the Apples are now PC standard. The best thing Apple did was let you buy PC100 RAM for cheap, and have it work.
Looking for freelance Actionscript (Flash/Flex) or ColdFusion work and/or freelance developers. Email me, put Slashdot
>Wouldn't it be cheaper to do this with a bunch of
>PC's in stead? Not as pretty but it's someones
>tax dollars that pay for these...
This myth has gone on enough. Every TCO study around shows that Macs are cheaper than PCs. Most of the TCO of a computer is the support time and training, NOT the purchase price of the hardware/software.
Macs require less support staff, for more computers. If I recall correctly, Intel has one IT person for every 30 computers. Except in their graphics dept, where they have 1 support person for all 300 Macs.
Because the Mac's interface is more standard across most applications, when you train on one app, that training can be utilized in other apps. This means that you get better training than if you have to train for every app.
It was recently figured that the difference averages between $400-1000 per machine, per year.
If you want to save your company $, you'd switch to Macs, but noone will do that, because it would cost (at least) half of your IT support their jobs.
Well, i'm not a native english speaker, thus i have to look up the word cluster in the dictionary, to get its real meaning. And this tells me, a cluster is nothing more then an amassment of something, and in this case of apples :)
..." use foreign words always in the wrong sense ... - 'cos everybody's doing so.
OTOH there is some saying here at my place: "Fremdwörter immer falsch verwenden
:-)
How can they say that this is the worlds fastest Mac cluster when they haven't benchmarked my cluster of 2,000 derelict Mac 68020s strung together with Appletalk and Duct Tape??
"Glory is fleeting, but obscurity is forever." - Napoleon Bonaparte
Keeping
Why the hell they're using Fast Ethernet is beyond me. If you're going to shell out the money for 16 dual G4s, why wouldn't you spend the extra money and install gigabit networking? Or does 100Mbps suit the purpose fine.
- Ando
You are the weakest link, goodbye.
What about calling it an orchard?
I donate all spillover Karma to the charity of my choice... Ada was still a babe despite what people may say...
Wouldn't it be cheaper to do this with a bunch of PC's in stead? Not as pretty but it's someones tax dollars that pay for these...
Well, as the other replies said:
It took them considerably less than a day.
That is actually an achievement. As far as I can tell, the costs of personnel dwarf the costs of hardware, where an engineer costs upwards of $800 a day, and engineers sitting idle costs as much as engineers busy doing work, so the cost deltas of the Macs is offset by the speed and effectiveness of setting up and maintaining the array.
In the long run, it's not the fixed, sunk costs that a PC represents that is meaningful, it's the daily ongoing performance, productivity, and maintanence costs that mean anything, especially in a place that runs plasma tokamaks at $125k a day...
Geek dating!
GPL Deconstructed
- Distributed Lock Manager: locks can be simulated with certain types of MPI messages.
- Cluster-wide File System: use MPI to pass data back and forth between nodes, including instructions on where to write the data. Not only is there a cluster-wide file system, it's a customizeable cluster-wide file system!
- Process control: Scyld's bproc allows all processes to appear as if they are running on the master node. You can also move processes between nodes transparently with this.
- Connection manager: Scyld provides this to some degree. You can do remote shutdown/startup of nodes or groups of nodes.
- Shared System Disk: well, nodes bootstrap from thier own drives, but download a new kernal image from the master node on bootup. They also pull down libraries from the masternode on bootup.
- Single security and management domain: permissions are the same on the slave nodes as they are on the master. But the slave nodes are truely compute nodes, and permissions there matter little, except for data files.
- Cluster wide process control: you said it. Beowulfs do it.
- Mixed Architecture: what you can do depends on your cluster. For a mixed Mac/Alpha/x86 cluster, you'd have to have different executables for each node. I'm sure there's Beowulf software that lets you do this, but for ours we don't need it (all Athlons.)
- Rolling Upgrade Support: yup. Acually, with Scyld, if you reboot a node, it will come back up with the newest configuration, imaged off the master node.
- Parallel IO support: simulated/managed through MPI pretty easily. Set up "IO" nodes and let them handle it.
- Interconnect failover: networking on a Beowulf is up to you. We use a high-performance switch and some channel bonding. It can be done.
- High-end scaling: Beowulf? "OK?" Have you ever heard of ASCI-Red?
- Load Balancing: ours does round robin scheduling of jobs, but it usually doesn't utilize the higher number nodes unless you run a job that requests a large number of nodes. We wish for better control, but this works pretty well.
- Cluster Alias: yeah, what you said.
:)
Anyway, the URL for our machine is http://www.cs.uidaho.edu/thecollective. It's up to 64 nodes right now, even though I don't think the site mentions that."While I think this is a sweet application of technology, it continuously galls me to see the misuse of the term cluster when applied to distributed parallel computing arrays such as AppleSeed or Beowulf. These are not clusters! There is no distributed lock manager present, no hardware-level device sharing (as opposed to software-level file sharing a la AFT, SMB, or NFS), no means of transparently starting, controlling, and stopping any process (not just those written to a custom API) on any node from any other node. While an excellant and usefull technology, this has a long way to go before it could be considered a true cluster..."
Blue skies... Barthie burgers... girls.
Were they really the first? What? to cluster OS X, or to use MAC Hardware in a cluster? I agree it's a neat deal, buy I'd also have to agree they probably did it for the press.
--
--Got Lists? | Top 95 Star Wars Line
If you used P3/ P4/ Athlon machines, then you'd have a much cheaper cluster.
Even Slashdot wants to hide some things
Amoeba which is now freely available for anybody to use.