Slashdot Mirror
Cringely Wants A Supercomputer in Every Garage
Nate LaCourse writes: "Real good one from Cringely this month. It's on building his own supercomputer, but with some twists." You'll probably also want to check out the KLAT2 homepage to learn more about their Flat Neighborhood Network. And since KLAT2 has been around for nearly a year (check out the poster on this page!), perhaps a 3rd generation is in the works?
Am I the only one to spot the "The Day The Earth Stood Still" reference here?
how sad...
Make sure it has a red dot and says things like "Dave, what are you doing Dave?" Can't wait for mine!
Satanists get good grades too...suspiciously good grades
It seems in the article that an extensive amount of calculations was necessary to design the network. Ironically, they needed a supercomputer to design a supercomputer.
It is really cool considering that $6,000 is now enough to take on massive projects. How many months would this machine take to render Final Fantasy: The Spirits Within?
I just love the idea of having a little super computer (and not just buying a Mac cube which claims to be a super computer, or a Dreamcast which can't be sold to Iraq because it qualifies as a supercomputer). Making a super computer must qualify as one of the ultimate hacks, a combination of technical skill, imagination, and pure unadulterated tech balls. This seems like one of those projects that I would do if I had the spare 6k needed. And oh yah, imagine a Beowulf cluster of these!
Sleep is for the weak!
Haveing a super computer would be great, imagine have the possibility of being able to do your own DNA Research at home. Or you could just get like a 1 gig Vid card and play an awsome game of UT!
Barada Ni*cough cough*
Well, yeah if UT is threaded.
I realize you were joking, but seriously, do any current games use SMP to their advantage? What about running games on a MOSIX cluster? Has anyone tried it?
I've had enough abrasive sigs. Kittens are cute and fuzzy.
A story that beowulf cluster posts will be relevant to!
Remember, there were no nuclear weapons before women were allowed to vote.
This is a very interesting concept that he is putting forth, but at the same time, how many geeks out there are going to really make use of such a clustering farm? Not everyone I know does video compression projects, and it would seem kinda prohibitive for a black-hat to set one up to break encryption codes. Could someone please tell this naive soul what useful everyday application all these CPU cycles could be used for? (if you say SETI@Home, I am going to bitch-slap you)
Secondly, UWB seems to be the holy grail of wireless networking, yes, however is this something that the agencies of the world are going to let out of the bag so easily as he says, I can think of the CIA and the NSA having a few choice words about such "undetectable signals" being used by commonfolk after September 11th...
Just my two cents
ZDNet has an article of HP building a supercomputer like this as well, called the "I-Cluster." It has 225 networked computers running Linux Mandrake (so changes could be easily made) on 733 MHZ out of the box PCs. The only catch is that is is slightly more expensive- $210,000 (minus network cabling). On the other hand, they plan to release the open source tools they made as well, so that people can repeat this.
"I have not failed. I've simply found 10,000 ways that won't work." --Thomas Edison
I always think that it isn't worth to waste the valuable garage space on my second-hand japanese car, which worth no more than $1,000.
Now it is used to place a $41,000 supercomputer! Ph43r m3!!
but then, I wouldn't allow anyone driving a car into my garage(WATCH THAT NETWORK CABLES ON THE GROUND!), so should I build another garage for my real cars?....
Those are some interesting ideas.
Now how about organizing them before publishing them? Call me pre-postmodern (and I'm still in my twenties), but I tend to learn more from a coherently-organized message than from a random jumble of statistics and facts. Cringely jumps from a detailed description of the KLAT2 and its innovative networking technology to a brief description of UWB. And then it's over.
Maybe I'm missing something.
"First you gotta do the truffle shuffle."
Now what am I going to do with the extra computational power that I created?
Running Super-SETI at home, claiming to be the greatest contributor when they really find ET?
Running Super-Quake with all the transparant cheat-code on without a slight jitter?
Rendering MSN frontpage in less than a second, with Mozilla?
Any better idea?
Well, not to be one of those stick in the mud 'Read the %($#ING article' type people, KLAT2 is a reference to The Day The Earth Stood Still. Had you looked at the articles in question (particularly, the KLAT2 page) you would have discovered that indeed, they were intending the reference. Heck, go check it out - the poster they made up for it is worth the look! :-)
Davis Ray Sickmon, Jr - looking for something to read? Check out my three free novels at MidnightRyder.org
Ignoring for the moment the pressing question of why any individual would actually need a supercomputer, I prefer to revel in the fact that it is even possible to have one.
What good is a supercomputer in your garage if you do not use it to maximize garage-holder value? If you provide supercomputer habitat for the progeny and supercomputer embodiment of the JavaScript AI Mind, which has also been coded in Forth as Mind.Forth Robot AI, then your home-sweet-home garage will be a major waystation on the road to the Technological Singularity.
Just as the Shroedinger Equations for atomic bombs and such were developed seventy-five years ago when Erwin Schroedinger spent his 1926 Christmas vacation holed up in the Swiss Alps and working out a few mathematical formulas that shook the world, nowadays over the 2001 Yuletide there have been the first stirrings of True AI in the JavaScript AI Mind, which any garage tinkerer may adapt for either 'pert near all-powerful supercomputer AI or a killer-app if not killer robot.
Following in the footsteps of the giants who created Visual Basic Mind.VB and Java-based Mind.JAVA, be the first on your block to create the supercomputer-based Garage-Mind.
Quake 3 Arena can use multiple processors (set r_smp to 1). However I havent found it to be too stable....
Other Q3A based games may also use this (RtcW does, but on the demo at least, its _Very_ crashy).
I thought it was 100base-T? Am I missing something?
Here's the qoute "And fast Ethernet (10base-100) costs about three percent of gigabit Ethernet on a per-card basis, so using four cards per PC still saves 88 percent."
Google results for 10base-100
- Results 1 - 10 of about 81.
Google results for 100base-T
- Results 1 - 10 of about 64,800
Zoid.com
where do we get um? where do they go, it seems u have enough then u run out! does ne1 know of a site that sells um?
I know someone that would love to see this in a nastran enviroment. Only if MSC could be less tight with there linux licenses...
I wonder if he's referring to Stahn Mooney's wife from Rudy Rucker's *ware novels...
Proteus' Child
Doko ni datte; hito wa, tsunagette iru.
Maybe they could set things up so that ALL his articles hit the main page as soon as he posts them.
If this were the case he could put a "discuss this article" link on his page and simply link to /.
Lasers Controlled Games!
Sheesh...!
You're using her as bait, Master!
Still though, after having to wallow through Cringley's painful lack of comprehension of basic technical knowledge, reading the ArsTechnica piece again was quite refreshing.
Is your company running tools written by ma
Through Google I found the UWBWG, and there's lots of detailed papers at Aetherwire. Interesting reading.
What do you think of MusicCity now?
... but it isn't going to happen any time soon :)
...
Unless we redefine "supercomputer" to mean "something that would look awesome to people 20 years ago."
Heck, the Trash-80's would've been a marvel in the ENIAC's time. It's like if everyone was "SuperMan" -- would we call NormalGuy a superhero for getting rid of those Kryptonite fiends for us?
Wait, what on earth was I talking about again?
Nevermind.
Just search in the Slashdot archives. I'm sure you will find MANY more references substantiating Mr. King's untimely demise. :p
Cringely is completely missing the point. KLAT2 uses multiple routes and switches, not channel bonding. And what the project contributes is not the basic idea of using multiple network interfaces (which is decades old), but a specific approach: using genetic algorithms to optimize the network topology. More traditionally, such clusters have used manually designed topologies with known performance bounds.
Speaking as someone who, yes, has actually worked with the big iron...
Why bother. Remember, Moore's Law is still in effect. Recently, we've hit the point in the curve where supercomputers are no longer needed, nor cost-effective. That is, the time it takes for the industry to deliver a far superior product has eclipsed the average lifespan of your typical supercomputer.
We're living in an age where a single graphing calculator you can buy at Walgreens has more horsepower under the hood than what got us to the moon 30 years ago. Your $2700 PC will be worth $150 within 3 years.
Having a supercomputer in every garage makes about as much sense as taking a rocket fuel-powered dragster to the supermarket for a gallon of milk.
Cheers,
Bowie J. Poag
I'd rather have a superMODEL in every garage.
7 November 2006: The day Americans realized corruption and incompetence weren't addressing 11 September 2001
Are you saying that the reports of his death have been greatly exaggerated? :0
"The economy is tanking, we're at war, the national and international situations simply cry out for escape, denial, and delusion."
I can recommend some nice psychotropic drugs for you. You could put on a white coat with magical powers and stay in a fortress of solitude.
Have you played q3 or UT on a sweet athalon?
Can you even tell the difference in game play between say a 900mhz and a 1.4?
mostly, it seems to be the the video card that makes the difference, and the ram, not the processor.
Of course, heat *is* an issue... but imagine a half inch between each layer, you would rack mount them at a slight angle and use heat convection to pull up air, a chimney effect...
:)
$1,499 for a 600MHz iBook, 20 of these would cost ~$30k, but you couldn't use the channel bonding concept, unfortunately. You'd be stuck with 100bT, which would probably get swamped with any real work in a 4 iBook per switch, 6 switch topology... without even trying to minimize latency.
20 iBooks would also take up about
8x9.1x11.2 per stack, so all 5 stacks would take up about 40 inches in space... You could stick these next to a desk or bed and use it as an end table! Okay, that'd be a tall end table...
$2,999 for a 667MHz ToBook, 20 of these would cost ~$60k, but these *are* Gigabit capable! In a similar topology, or perhaps because of prices for Gigabit switches, you might as well use one switch. Who knows?
Of course heat is even more of an issue, but give n the same space as the iBooks, there's a whole extra half inch of space available to the TiBook!
40x9.4x13.5 inches! It would even make a good space heater!
Okay, okay, I know, it's damn expensive. But... consider, how much is a 20 CPU machine from HP or IBM? I know, I know, they tackle different uses, like reliability, uptime, IO throughput, etc. A 4way 680 pServer from IBM is $220k, from their own website
Damn... I wonder when Apple is going to release a thin rackmount slab server?
GPL Deconstructed
The costs of a clustering setup go well beyond the initial hardware. At the level that Cringely is building (with only 6 machines), it may not be a huge problem, but running KLAT2 will cost you some dough just for the power.
A couple years ago I made a dumb mistake and bought a saltwater reef tank without realizing that it would end up costing me $150/mo. in electricity bills (it ain't cheap running 4000+ watts in lights and pumps 18 hours a day). I'm sure running 66 machines 24 hours a day ain't cheap either.
...and it still won't run Quicken. Enjoy!
Cringely Wants A Supercomputer in Every Garage
Is he buying?
Also, if everybody had a supercomputer in their garage, they would no longer be so "super."
...you'll be lookin' at a whole lotta Kentucky Fried Penguin!
Anybody want a peanut?
My god, but you're a tedious freaking idiot.
Dr. Dietz used to teach at Purdue, and I had the good fortune to take a compiler course taught by him. On the first day, when introducing himself, he came to the part where he was describing how to get into contact with him. When giving out his phone number (at Purdue, on-campus numbers were 5 digits long) he mentioed that his phone number was "GEEKS". He added, "No, I didn't ask for GEEKS, but when I figured it out, I thought it was pretty cool."
Needless to say, it was a pretty cool course.
Can you imagine a beowulf cluster of these supercomputers??!?!
...I'll get me hat...
"Reactionaries must be deprived of the right to voice their opinions; only the people have that right." - Mao
25 dual p650s in my home office ... when I crank it all the way up I come in at somewhere in the 50-100 range on the dnet rc64 dailies. Sadly the original reason I built it has evaporated and with the current cost of CA power I just have a fraction running
tis weird... just at 3:00 in the morning i was thinking of a 3-4 boxed p2/128mb ram beowulf cluster instead of a pentium4/640mbram...
<disclaimer>I know little about 'big iron'</disclaimer>
But isn't the point of these kind of projects to derive more computing power in a generic form, something useful to many situations?
Sure, my Athlon isn't too slow at the piddly little hobbyist 3d rendering stuff I play with, but what if I suddenly get grandiose dreams of 3D worlds, wouldn't it be nice if I could divert the down payment for a house and move myself a year or two farther along Moore's timeline?
I can think of some small business applications where a nice quick video compression would be nice, especially if the hardware and software were all generic enough to buy off the shelf without a serious outlay of cash. Granted, there are very nice and very fast hardware codecs but then what if that same small business wanted to render some 3D along with that video stream? Or I'm working for them and get permission to render my VR opus overnight?
What about applications that could be enabled by cheap and standardized GFLOPs? If you can't think of any you're not thinking hard enough.
Bleh!
Clusters are great for embarassingly parallel applications (ie ones that have threads which don't communicate with each other much. This includes things like SETI@home and batch rendering of images. What they don't compare on is applications that communicate a lot like nuclear physics simulations. This is not to say that that will never change in the future, but for the time being it's still true.
Last, and certainly not least, real supercomputers have memory bandwidth that can match the speed of the processor. A Cray or an SGI Origin has an absolutely massive amount of bandwith from the processor to local memory compared to a PC. That allwos a traditional supercomputer to actually *achieve* the fantastic peak performance numbers. On many applications, the working sets are huge and don't fit in cache so you end up relying on memory being fast. On a PC, it's not and I've heard from sources I consider reliable (though I have no actual numbers to back this up so it may be rumor only) that one large cluster site sees around 10% or less of peak on a cluster for a nuclear physics simulation, whereas, on a vector Cray, you can hit ~80% of peak. This means that the cluster has to be 8 times more powerful and when you start multiplying the costs by 8, they start looking like the same price as a real supercomputer.
So my point is that building a real supercomputer does not mean grabbing a bunch of off-the-shelf components, slapping them together with a decent network and running Beowulf (or a similar product).
Go Badgers! -- #include "std/disclaimer.h"
A housemate of mine and I decided that we wanted to build a pathetic little Supercomputer out of the various PCs laying around in our little Geek House. We've decided to give MOSIX a run. It sounds like a fairy tale solution...especially when it comes to automatic process migration node to node. Anybody here have any positive experiences or harsh words regarding this?
While he's giving things away...
...can I have a pony?
------
Today's Top Deals
I remember reading a book on robotics and one of the first talking ones was trained to say what the trainer was saying. The trainer said "You Talk" and it said "Klatu" (you talk backwards)
Must have had a screwed up for loop somewhere.
Don't know if the its the same reference, but it reminded me.
They should recalculate all this if they have Gigabit LAN cards. You can get them from buy.com for about $57. I have used them at home to make really fast point-to-point links. Also, even though the switches with all gigabit ports start at around $700 for $200 you can get a switch with 8 10/100 ports and 1 gigabit port. That should add some interestings properties to the network.
"If I can see farther it is because I am surrounded by dwarves." -- Murray Gell-Mann
The FCC is being very cautious about mass-market UWB products. Since these things blither over a gigahertz or so of spectrum, they overlap with other services. At low power, a few of these things are probably OK, but in bulk, there could be trouble. The concern is that mass deployment could wipe out other services in congested areas.
Wasn't the ps2 banned in china becasue they were afraid that the goverment would use them for atom bomb simulations? (or was that a rumor passed by sony to create more hype) It seems from mister X's article any one can build a supercomputer with some knowledege and some dough, what's stopping the "Rouge Nations" from designing deadly atomic weapons(if they haven't allready done so) The ablity to process vast amounts of data is no long for a select few but brought to the masses, do you think there would be consequences?
Me and lunchbox here are going to kick your ass.
I want to see a Beo--...err, wait, never mind...
// file: mice.h
#include "frickin_lasers.h"
...a single-CPU version of this!
From the PBS article:
"The solution was to put more cheap Ethernet cards in each PC, and then use "channel bonding" to make them all look like a single faster card"
From KLAT2 FAQ
"Every NIC in every PC has a unique MAC address (and potentially unique IP address) -- i.e., this is not channel bonding."
FNN is totally different and in many cases more suited for this app than simple channel bonding.
One thing I did wonder about was. Why the floppy drive? You can get netboot cards very cheap... And you'd only need one per system. Just one less mechanical thing to fail. Plus the node would come up much faster. PXE or even BOOTP/DHCP boot would be fine.
Also I kind of wonder about commodity Realtek cards. I'm sure Realtek makes a fine chipset, but most vendors who use Realtek chipsets really skimp on the rest of the card. You can get 3com or Intel Pro/100 multipacks almost for the same price as the realtek cards sell for off of the shelf.
Ok rant over... Flame away.
the article forgets development costs; development of large parallel applications cannot be done on a MUCH smaller one.
my guess is the best strategy is something like:
1) devellop a prototype on a tiny cluster
2) buy (or rent) a medium-size, medium-speed cluster and iron out the network-related and paralelism-related problems.
3) then look at moore's law and the hardware roadmaps, and decide when it is best to buy the super-computer for actually solving your large-scale problem (which is what parent-comment discusses).
Working for necessity's mother.
Yeah I would have really liked to see them do a 6400x4800 full screen rendering of a scene in a parallel optimized version of POVRAY. At least it would be something I could understand as far as speed goes.
Am i the onlyone who see's the posibilites of this
Ya, you could finally run that spell checker you've been meaning to!!
I like this kind of home supercomputers some examples here of course, linux powered
Look up gigabit on Pricewatch, there's 64bit pci Ark cards ( based on Natsemi I think) for less than $60.
The most expensive part of gigabit is now the hub/switch.
The illuminati here are missing the point of the experiment. It is to obtain large quantities of processing power for math heavy processes. This is not for a render farm.
This guy is going to building a second one for a neighbor and transmit the first commercial UWB video transmission outside of a lab.
I beleive the US goverment clasifies any system that is capible of over 1gigaflop as a supercomputer. IIRC, The G4, P4 and Athlon all deleiver at least a gigaflop....So, many of us DO have supercomputers in our garage. (I keep mine in my bedroom)
The opinions in this post are ficticious. Any similarity to actual opinions, real or imagined, is purely coincidental.
How can 4 bonded 10/100 cards provide better throughput than a gigabit card? I can understand them being cheaper, but the idea of them being faster is hard for me to grasp.
The way I see it, 4 100 megabit cards at MOST would create a 400 megabit pipe. Wouldn't bonding add overhead to it to make it even less than 400 megabits?
Would somebody mind explaining how this works?
"Derp de derp."
Someone set him up the bomb.
When I read the title, I had visions of actual homebrew *supercomputers*...something along the lines of Euclid from Pi. Break out the soldering irons and damn the torpedoes! Yes, Beowulf technology is great. Genetic algorithms and channel bonding and QNX are nice touches, but Beowulf clusters are fairly common, even for ordinary people...an article about NEW uses for the things would have been nice, other than that, just shut up and build one! Soon to be taking my own advice...even have an app in the works for it. Still a nice project with a half-decent writeup, but it's been done.
How about it, folks? Homebrew big iron? Where would one even begin? Food for thought, at least.
"These people look deep within my soul and assign me a number based on the order in which I joined" --Homer re:
One: Because we can.
Two: Ever seen the stuff they run on supercomputers today? Simulating a supernova for 1 nano second can take a month of CPU time on some of the world's fastest supercomputers. Oh, its still very nessesary. If the past is any indication of the future, we will always need blazing fast machines to push the limits in the scientific world.
I assume you mean big iron as in mainframe, which is NOT a supercomputer by any means. Mainframes do the work that runs this world, supercomputers help us discover what we'll do in tommorow's world. They are very different worlds.
--Nuintari
slashdot : where an opinion can be wrong.
Klatu Barrada Nickto
----- Whats wrong with this picture? http://www.revoh.org:1234/whatswrong
A supercomputer in every garage is something that is too expensive and useless most of the time, but a supercomputer in the neighborhood could be a relalistic and usefull idea. We have seen on slashdot that is it possible and not too expensive to make a neighborhood fiber lan. Most people do not need much processing power all the time they need a high peek processing power for short moments.
Most user application software will not be able to use divided the load over the cluster but there would be many applications running at a time, so the load would be spread over the cluster without special application software. People could keep there old PCs and turn them into X terminals or use vnc to connect to the cluster.
$41,000 is only $205 for everyone if 200 people use the cluster.
Jan
"(the operating system) will be QNX, a real time OS that supports massive parallelism and has very low overhead. QNX is fast! QNX is also Posix compliant, so there is lots of software that almost works under it."
If you're looking for software that almost works, I know of an OS that might fit your needs. You're not going to hook this thing up to the Internet, though, are you?
"I-Cluster".... but how many colors does it come in?
Duh folks, UWB is just another term for spread spectrum.
Hello? Morons.
By definition only the fastest devices are supercomputers. These days that is about a teraflop. Thta includes the US DOE ASCI series and the announced installation of the Blue Storm and Blue Gene IBM computers. Ten gigaflop computers a dime a dozen and a hundred gigaflops not so rare.
Uh, Cringely, wouldn't creating the thing and then using it as the subject of an article for the company that employs you count as a commercial purpose?
m00.
Yes, these systems are not sometimes the best for handling vectorizable jobs, but they are so inexpensive compared to the old specialized hardware that it is easier to waste cycles than build special hardware.
As to memory bandwidth. Modern CPU caches make the question nearly moot.
If all of this were not true, then people wouldn't be building clusters and the majority of the top500 list wouldn't be dominated by clusters. Instead there are 3 traditional architecture machines in the top 20. This is the reason that Cray (etal) no longer dominates the marketplace... commodity systems have overtaken nearly all of the specialized hardware world.
-- Multics
Also think about the possible misuse of the abilities to perform this amount of calculations. Improving nucular warhead blast yields, geneticly engeneering a new super virus taht could wipe out the planet. To me this scares me more then anything else. I personally dont need a supercomputer to chat on IRC, post to /. and type up and compile my projects.
I will bend your mind with my spoon
It's spelled Lamborghini.
/.: why the hell am I here?
It wouldn't help so long as the CPUs couldn't utilize the gigabit bandwidth. Swap out the 100 megabit lan cards for the cheapo gigabit ones for slightly more money- I think you'll find that this cluster's still starved for bandwidth.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
Well, don't we all :)
----- Whats wrong with this picture? http://www.revoh.org:1234/whatswrong
Most of what Cringely said may have been true a year or so ago but it isn't now. For instance, if he is using Athlon XPs, why would he move FP code to 3DNow! instead of SSE? And there are a host of competing cheap interconnects now, especially if you can avoid TCP/IP. But if you can't, there's always IP-over-FireWire... M$ has had that running at 400Mbps for years. I don't think you can get into UWB for $6000. See the Linux Clustering Info Center and Extreme-Linux.com.
You said: "The answer is that Cringely's (sp?) collection of processors is not a real supercomputer for the kinds of applications that are associated with traditional machines. ... Clusters are great for embarassingly parallel applications (ie ones that have threads which don't communicate with each other much. This includes things like SETI@home and batch rendering of images."
Cringley said: "Beyond using it to heat my office, I plan to keep the supercomputer busy with a video compression project I'm doing as well as further experiments in wireless communication."
Sounds like he's using the right tool for the job, then. I think he's using the term "supercomputer" to refer to a machine that is many, many times more powerful than the PC that a typical user would have sitting on his/her desk. By this definition, my PIII-600 is a supercomputer compared to the 486SX-25 I started with in 1992.
The man who does not read good books has no advantage over the man who cannot read them. - Mark Twain
Lets also not forget the little matter of needing software that can take advantage of parallelism. Not being familiar with QNX I don't know if it can schedule threads or processes across networked computers (my guess would be no), but tools for parallel execution (especially compilers and efficient networking libraries) are still hard to write and generally expensive. Something like MPI can probably be gotten for free these days, but then you're back coding Fortran or C or something, always with a watchful eye on parallel performance. That takes experience and time, both of which aren't cheap. Also, if any communication needs to hapen between processors, network latency usually tends to be a bigger problem than network bandwidth, unless large amounts of data are transferred (in which case good performance may be impossible due to a low CPU/communication ratio).
d00d! Imagine a Beowulf cluster of DREAMCASTS!!!
I don't understand why this guy's so excited about 64 gigaflops for $40,000. Consider the following
6 Dual 800Mhz G4 PowerMacs (running your choice of Darwin or Linux)
Processing Power: 70.8 gigaflops (11.8 each machine)
RAM included: 1.5 GB (heck, I already have 1GB in my PowerBook G4)
Hard Disk Space: 480 GB
Extras: Each machine has one 4x AGP slot, four 64-bit PCI slots, an NVIDIA GeForce 2 MX card (64MB), Gigabit Ethernet and a 56K modem you can just take out and hang on your Christmas Tree or something
Cost: $21,000 ($3,500 from store.apple.com)
Ebay off the CD-R/DVD-R Drivers for $350 each to save $2,100
Total cost: $18,900
Tip: If you wait until after MacWorld in January (where Apple will most likely introduce a line of faster PowerMacs) you can either get more gigaflops for the same price or lower the price of 64 gigaflops to around $13,000.
What kind of an idiot goes out and spends over $40,000 on something less powerful than he could get for $13,000? And people complain that Apple hardware is expensive... morons...
How much do embedded processor like the ARM cost each? Why can't a supercomputer be built using 256 or 1024 chips on a single board? I'd bet that the result would be cheaper and much faster than clusters of networked PC's - due to greatly (by thousands of times at least) reduced communications costs.
>But what do you expect from someone who practices identity theft [wired.com]?
How exactly do you steal the identity of someone who never existed? The man we know as Robert X. Cringly was the Infoworld Cringly for 8 years! I'd say he pretty much defined who that Cringly was (or is today, I don't read Infoworld.) Saying he practices identity theft would be a valid argument if the Infoworld Cringly was someone else and he had just appropriated the name for use on PBS, but he didn't. He built up the Infoworld Cringly and so I believe he has a right to go on with the persona he's used for all this time.
man RTFM
No manual entry for RTFM.
I have a TiBook. Apple *could* do away with the screen, keyboard, and speakers, and replace the CD-ROM slot with a ram-bay.
:)
Not only could you hook them together using gigabit ethernet, you could take advantage of the firewire port as well, perhaps chaining them together with some sort of SAN, though you are still limited by the ~50MBps, though perhaps that's not useless, I don't know.
Still, with the ram bay you could up the memory from 1GB to something crazy, like 16GB. The battery is useful as a backup-emergency device, allowing the slab to run for about 4 hours in case of emergency (woo!).
You could even concievably netboot the thing, since OS X allows for that, right? Minimize the hard drive or get rid of it altogether... you could seriously make a slab about the size of 1/2" by 8" by 8" I suspect
GPL Deconstructed
Quake 3 supports SMP, try /r_smp 1
for me, it really speeds things up
I can get a clone Realtek card for $10 (or less sometimes) retail. Most other cards are at least $20. I can also pick up a floppy for $1 a piece used.
Not sure how much you could get the realteks w/rom for bulk but you would have to call around
make Linux, not Microsoft. sin(beast) = -0.809016994374947424102293417182819
Can this be done with sparc's?
linuxhelpnetwork.homelinux.net
1) heat in garage in winter
2) Top 10 in Seti@home
3) Porno-ize you favorite anime (Final Fantasy anyone?)
4) Why are you reading this? I thought you were doing #3
make Linux, not Microsoft. sin(beast) = -0.809016994374947424102293417182819
Fast ethernet is good enough for pretty big range of well-written parallel applications. In general, the ratio of CPU power/communication has to be proportional to nonlinearity/nonlocality ratio of the problem the computer is designed for. Usually, for partial differential equations, explicit timestepping algorithms put little strain on communications. OTOH, once you go to implicit transport equations of any kind, you need all communications you can get.
:(
And, memory bandwidth is a bitch, I agree. Those dual athlons really have memory bottleneck
It has twice the bandwidth of older PC hardware. This should meet your needs much better, and is still a lot cheaper than specialized busses.
And Quad Data Rate (QDR) memory is on the way.
I do agree that the internal busses of modern PC's are woefully inadequate. Many of them still support ISA for some sadistic reason. Probably need it for the floppy controller.
I want to see very high speed serial data busses between every component hooked to an internal switch, so that any two components can directly communicate with each other without having to go through the CPU bottle neck.
This would also allow you to add as many memory modules, cpu modules, hard drive modules and the like as you have switch connections in your box. I think that motherboards are on the way out.
I think that so far in the year 2001 Stephen King has died at least four times. It's one of those things that keeps popping up like those damn gophers in the arcade machine where you hit em with a mallet (whatever the hell it's called).
-
Steve's Computer Service, Hobbs, NM
Not to sound like an ass, but this is the best article I've read on Slashdot for quite some time. So when I want more information on the subject I was hoping to find some intelligent or at least funny conversation and most of the posts are crap, just when my 3 day moderating period has ended. I wish some people would just read the linked articles before posting!
if common sense was common, wouldn't everyone have it?
Uh, Cringely, wouldn't creating the thing and then using it as the subject of an article for the company that employs you count as a commercial purpose?
You don't really expect QNX to bitch about a little free advertising do you?..
Bitch, my fp was on topic!? aren't we still talking about beowulf clusters? isn't that relevant to supercomputers like Cringley was referring to...you dumb fux...HOW DARE YOU MOD ME DOWN AS OFFTOPIC!
When you are comparing costs for construction, don't forget the essentials.
16 Pizzas for student helpers @ $10
4 Cases of soda student helpers @ $7
Total: $188
Two useful definitions that explain why KLAT2 was built as it was:
The same argument applies for latency: single switch for 100Mb/s FNN versus multiple switch hops for Gb/s.
It's interesting reading the replies here. It's as if all clusters are only defined by CPU and speed of the NIC. Sorry, it's not that way at all.
A cluster should be designed to solve a specific problem. You have to do some math up front before making your choices on things like NIC's and CPU's. You can't apply a blanket solution. Some of the things to consider are:
. The size of the data being exchange
. Does the matrix you are computing fit into cache?
. Is the problem dependant on memory bandwidth?
. etc....
For example, for some problems, the amount of data exchanged fits well within a Jumbo Packet of a Gb Ethernet AND is not affected by latency.
Other problems are very dependant on latency and require things like a Dolphin or Myrinet card. These are not inexpensive items.
Still other problems require memory bandwidth and work well with systems like Alpha's.
So, if you are building a small cluster to run POVray at home, go with cheap. If you are trying to crack the human genome, you need racks upon racks of things like high-end Alpha's with Quadrics interconnects. It's all dependant on the problem you are trying to solve.
FWIW, I used to work at API and DEC/Compaq.
What's my Karma Mr. Burns? "Excellent"
I went to that movie on a date... and I liked it. A nice japanese cinema style movie.
FF:TSW and now that you can get a super computer for $6,000 is a little scary -- how long until companies buy the rights to peoples appearance. Then force them to get plastic surgery, and have computers make all the footage for the news anchors/commercials.
https://www.gnu.org/philosophy/free-sw.html
the new doom game is supposed to be and it will supposedly run uber great on a dual processor
Whose idea was it to put Windows servers on the net in the first place, anyway?
HAL is _Much_ more user friendly than quicken. Based on the model of Checkbook ledgers? *shudders* I prefer a computer with a death wish for me than a program that wants me to keep a checkbook ledger.
He did bring it on himself though, by writing under a pen name to see if he 'still had it' when writing horror. Ever since his 'pen name' was discovered, and 'died' in a tragic accident when his tie was caught in his typwriter. Ever since then Stephen King has been dying several times a year. It's almost as bad as those 'forward this to 8 friends and get $50 from (insert famous company or person here)'
You know the next virus should exploit that vulnerability. "forward this e-mail to 8 friends and then run freemoney.exe to get $50." Just make the payload set the default gateway to 127.0.0.1 please.
I think one of the neatest ideas about having a supercomputer in your house is that there is nothing to do with it. Most people think this is a negative, having no problems to solve. I think it's fantabulous. Think of all the new problems to discover so you can have something to solve. Sure the lazy among us will go for Optimial Golum Rulers, digits of Pi, Chess, SETI, RC5-xxx, video rendering/compression/effects, or whatever.
IBM was going to/is build a monster of a machine Blue Gene. Biggest machine ever built. However they wanted a problem to show off their beast, they decided to look at protien folding. NOBODY was looking at protien folding using supercomputers to actually analyze them. It was a computer waiting for a problem big enough, and it fit.
I'd also like to note that though these computers being built have HUGE processing power they don't have the latency and bandwidth a lot of problems like Weather prediction. Unless you get a Cray, Superdome, or Regatta it's like towing a trailer with a piece of yarn. For certain problems the processing power is less important than the memory bandwidth.
Anyone who cannot cope with mathematics is not fully human.
http://aggregate.org/CDR/ , the Cluster Design Rules tool
You specify some characteristics of your application, your site (power and space), and budget; it presents the best designs taken from a design space of millions.
:)
TWO BYTES LEFT!