Supercomputer Breaks the $100/GFLOPS Barrier

Its about time by HanzoSan · 2003-08-23 02:22 · Score: -1, Troll

Now maybe we can find some use for these super computers and all these Gflops.

--
If you use Linux, please help development of Autopac

Re:Its about time by Anonymous Coward · 2003-08-23 02:27 · Score: -1, Troll

I'm sure it could be used to calculate how to better kill people... you know, just like (almost) all the other ones.
Re:Its about time by Anonymous Coward · 2003-08-23 02:33 · Score: 0

Sounds good to me, there's too many people around here already. Let's start with the yuppies that think the most important things in life are the status symbols they drive or that every blade of grass on their lawn is exactly the same height.
Re:Its about time by Toby+Studabaker · 2003-08-23 02:42 · Score: 1

Hey! Ho!
It's saturday evening and time to start drinking beer!
There's always use for Gflops. How about distributed DVD ripping, packing and then serving them to the people on the net?
Re:Its about time by Anonymous Coward · 2003-08-23 03:04 · Score: 1

I am a grad student at UK Computer Science dept...

They are working on projects to use this type of power ... see http://www.metaverselab.org

My favorite is the hooking together 16+ cheap ($2000 or so) projectors together adhoc to build a display that covers walls/floor, and combining that with head tracking and video cameras that look for shadows so that other projectors can fill in! This needs a lot of GFLOPS!
Re:Its about time by Anonymous Coward · 2003-08-23 04:02 · Score: 0

Huh? Isn't the obvious solution to use rear displays?
Re:Its about time by Anonymous Coward · 2003-08-23 04:56 · Score: 0

For the record, the Metaverse Lab and Hank Dietz have nothing to do with each other.

I can't tell by this poster's text if he means that they do or not, or is just talking about the University of Kentucky's CS department in general.
Re:Its about time by Anonymous Coward · 2003-08-23 05:00 · Score: 0

that's not always feasible, it's often more expensive, and there are problems with back-projection that are not present with front-projection

if you're back-projecting, you better have a nice surface that allows light through it (walls won't allow this, while front-projection is fine for walls)

there's also the "hot spot" problem with back projection where if your eye/camera/viewing device/etc is looking directly into one of the projectors, you'll see a hot spot...much much greater intensity from that angle

the biggest problem with front projection is occlusion by the observer...and the Metaverse Lab has solved this problem with shadow removal using multiple projectors...the unoccluded projector kicks in and shows the piece of the immersive display that was occluded by the observer

my anus by Anonymous Coward · 2003-08-23 02:22 · Score: -1, Flamebait

breaks wind

first post by Anonymous Coward · 2003-08-23 02:23 · Score: -1, Flamebait

first post

Wow! by fryguy451 · 2003-08-23 02:23 · Score: 5, Funny

Imagine a Beowu... errr... Oh..

Re:Wow! by wardomon · 2003-08-23 05:29 · Score: 1

A Beowulf cluster of Beowulf clusters? Where will it end?

--

- - - If the sun is a star, why can't I see it at night?
Re:Wow! by BSD+Yoda · 2003-08-25 04:19 · Score: 1

mmmMMMmm. Where BSD? If only this system running, better it would.

fp by Anonymous Coward · 2003-08-23 02:24 · Score: -1, Offtopic

fp

Let the Beowulf cluster jokes begin! by Anonymous Coward · 2003-08-23 02:24 · Score: 5, Funny

Note to moderators, Beowulf cluster jokes CANNOT be offtopic.

Imagine a Beowulf cluster of Beowulf cluster jokes!

Re:Let the Beowulf cluster jokes begin! by Anonymous Coward · 2003-08-23 02:27 · Score: -1, Flamebait

Thanks. Imagine a beowulf up your arse!

God damn I'm on topic today :)
Re:Let the Beowulf cluster jokes begin! by kasperd · 2003-08-23 03:28 · Score: 1

After reading about it on LWN earlier today, I tried to imagine this beowulf cluster computing how to build a larger beowulf cluster, or just compute how to improve it's own network. (Notice, they used a four node cluster to computer how to wire the network.)

--

Do you care about the security of your wireless mouse?
Re:Let the Beowulf cluster jokes begin! by blixel · 2003-08-23 03:45 · Score: 1

Note to moderators, Beowulf cluster jokes CANNOT be offtopic.

How about moderating them down for unoriginality then.
Re:Let the Beowulf cluster jokes begin! by ModernGeek · 2003-08-23 06:41 · Score: 1

Then this post shall be a node of the great slashdot beowulf cluster. Just imaging a beowulf cluster of beowulf cluster jokes!

--
Sig: I stole this sig.

Also I wonder by HanzoSan · 2003-08-23 02:25 · Score: 5, Interesting

How much electricity will these super computers use up?

All those wires, it looks like it takes up alot of juice.

--
If you use Linux, please help development of Autopac

Re:Also I wonder by jd · 2003-08-23 02:34 · Score: 1, Funny

That depends on how fast the students can operate the peddle-power generators.

--
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Re:Also I wonder by Shazow · 2003-08-23 02:35 · Score: 1

How much electricity will these super computers use up?
Heh, maybe that's what caused the recent blackout? :D

"Let there be li---*BBZZZZzttt*"

- shazow
Re:Also I wonder by HanzoSan · 2003-08-23 02:40 · Score: 1

You know I was thinking the same thing, besides the current excuse for the blackout is, "Well the tree hit the powerline and caused the country to black out"

I'm really hoping we did over stress the system because if the system is so sensitive that a tree rubbing against the powerline can black out half the country, well I guess we will have to get used to blackouts whenever a fly decides to land on the powerline.

--
If you use Linux, please help development of Autopac
Re:Also I wonder by lederhosen · 2003-08-23 02:43 · Score: 1

ca 100 boxes
ca 200 W/box

ca 20 kW
Re:Also I wonder by rusty0101 · 2003-08-23 03:24 · Score: 2, Informative

Per the FAQ on the site, the supercomputer draws 210A. Power requirements provide an yearly cost equivalent to the cost of the network equipment connecting the nodes.

210A at 120Vac via the power law comes to 25.2kw/hr. Tripple that to allow for cooling (It takes approx 2 watts of power to remove the heat generated by 1 watt of power usage) and you come to almost 76kw/hr. Take a look at your utility bill to come up with the hourly cost for electricity while this thing is on.

The equipment does not have cooling isolated from the rest of the building. As a result the cooling costs will ultimatly be absorbed by the operational cost of the building, and probably will not show up as a line item for the cost of this cluster.

-Rusty

p.s. more wire does not meed more juice, just more pathes for signals to follow.

--
You never know...
Re:Also I wonder by maraist · 2003-08-23 03:34 · Score: 1

Well, Athlons typically want 350W.
There were 128 machines.

Thats 44.8kW

Now take into account at 120V that's 373Amps rms. With all the surge supressors/power-strips, we're talking a serious amount of impedence (fwi, not resistance).

Not to mention, the typical circuit breaker clamps at 20amps. You'd have to have 36 separate circuits in a typical office environment (a circuit usually services several outlets).

To boot, the impedence at such high currents running off the same master power cables could cause serious power level fluxuations; especially at boot time.

Now this is a university; they're likely to have very clean power signals, and possibly each plug runs a seperate voltage-regulated power line. And they definately can sustain high power (look at the cables). But don't expect that sort of power center to run cheaply at your local ISP.

In a fit of coincidence, I've recently tasked out the problem of power distribution, and the solution I came up with was gutted DC AVR'd UPS's (with the DC->AC component gutted) that directly connect to the motherboard; bypassing the switching powersupply and any external drives. Given enough external cooling, you could even resort to passive heat-sinks (albeit large/copper/expensive). You could probably get the power dessipation down to 100W (75W for cpu, 25W for MB+passive periferals). It might be cheaper to get ahold of laptop style power batteries, but I don't know of an easy way of getting them off-the-shelf without just buying laptops outright (more expensive proposition)

The battery-operated systems gives you two things.. Lower total power dessipation, and greater power stability under non-uniform loads. Due to the AVR, it's also possible to use very long power cables to stretch across sides of a building (as a short term solution; avoiding rewiring the master power distribution)..

--
-Michael
Re:Also I wonder by jd · 2003-08-23 04:24 · Score: 1

There was work, at one point, on a single-stage high-voltage amplifier. The idea was to reduce the unwanted distortion by reducing the stages you needed to go through.

I think these guys need a way to tell if the computer has crashed or lost power. Y'know, UPS' have those mini alarms, but people aren't going to be around the computer all the time, and the UPS will only detect a power outage.

I think they need a watchdog circuit, linked to a 25.2 kilowatt amplifier and a suitable speaker. That way, no matter where they are in the US, they can tell if the computer has failed.

--
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Re:Also I wonder by b!arg · 2003-08-23 04:47 · Score: 1

Someone needs to work on their cable management skills too. One word: velcro

--

Everybody dies frustrated and sad and that is beautiful
Re:Also I wonder by lederhosen · 2003-08-23 04:53 · Score: 1

Athlons run well on 300W power supply.
I think the avarage power consumption is
way bellow that, considering that a normal
computer has a power hungry graphic card
several hard drives cd-rom and cd-burner.
Remember that the grading of the power
supply is the maximum it can deliver.
Re:Also I wonder by canajin56 · 2003-08-23 06:28 · Score: 1

Ummmm....my Athlon 750 system peaks at 150 Watts or so, and that includes the monitor, 2 hard drives, a dvd drive, and a cd burner. Without the monitor, it uses about 80 watts.
The reason Athlon and P4 system require a 300 watt supply is for when they are starting up.

--
ASCII stupid question, get a stupid ANSI
Re:Also I wonder by amorsen · 2003-08-23 06:59 · Score: 1

First of all your units are all screwed up. 25.2kW was right, forget the /hr thing. Second, I refuse to believe that you need 2W of electricity to move 1W of heat. Airconditioning seems to be more in the range of 1W of electricity needed to move 2W of heat. So let us say 40kW total, which in the silly units used for electricity billing comes to 350MWh/year.
Anyway, if you want to see stuff that really draws power, go look for the high energy physics stuff. Power cables that are liquid cooled through tubes in the copper...

--
Finally! A year of moderation! Ready for 2019?
Re:Also I wonder by Anonymous Coward · 2003-08-23 07:06 · Score: 0

Actually we were stressing the system.

It was the stress on the system that caused the wire to bend, which is when it touched the tree.

At least, that's what the History channel said on the special they had on last night.
Re:Also I wonder by Noah+Adler · 2003-08-23 07:52 · Score: 1

Not nearly as much as all the standalone boxes around campus. Power consumption has recently become a bigger issue here at UK, what with statewide (nationwide?) goverment budget troubles. Yet pretty much every professor (music performance, English, whatever) expects his own always-on computer in the office, plus a monitor (which KASY0's nodes don't require). Of course I can't really blame them, because computers are cheap, useful tools, but really, when you compare these power costs for the work they get done (sending e-mails mostly) to KASY0's power cost/work ratio, KASY0 seems much more reasonable to me.

KASY0 is doing (presumably) useful calculations that wouldn't otherwise be possible to do. Seems like it would be out of line to worry about it consuming too much power when so much more power is devoted to so much more mundane tasks.
Re:Also I wonder by LookSharp · 2003-08-23 08:39 · Score: 1

210A at 120Vac via the power law comes to 25.2kw/hr. Tripple that to allow for cooling (It takes approx 2 watts of power to remove the heat generated by 1 watt of power usage) and you come to almost 76kw/hr. Take a look at your utility bill to come up with the hourly cost for electricity while this thing is on.

On what planet? I cool my 60 watt or so Athlon XP 2000 using a 4 watt, 80mm fan. Add an 8 watt, 120mm fan on the intake that is WAY overkill, and a 4 watt PS exhaust fan, and I'm using 16 watts to cool a system that peaks around 280 watts power consumption. By your math, I should be using 528 watts cooling power for 264 watts used by the PC?
Re:Also I wonder by WoTG · 2003-08-23 10:34 · Score: 1

Which is fine for ONE PC. And, if you don't use air conditioning, then yes, a few watts in fans is all the cooling energy cost that you'll need. But try dissipating 25Kw. Your room and then house will heat up really quickly. Let's see, after the Big Blackout I think I saw an estimate that running a household oven is about 12Kw - so this cluster would be comparable to running two ovens 24/7. This heat has to go somewhere... and in most non-residential buildings, it'll be the air con. that has to get rid of it.

I don't know how efficient A/C's are, but it's definitely not going to be cheap.

I wonder if they could separate the computing room from the building's HVAC system and just run some fans to the outside on high speed. Enough air flow should be able to keep the computing room at more or less the ambient outside temperature. Probably uncomfortably warm during the summer, but it should be within the operating range of the PC's. How hot is a Kansas summer?
Re:Also I wonder by LookSharp · 2003-08-23 12:56 · Score: 1

OK I forgot about datacenter conditions. Sad considering that I work in a corporate datacenter with 3,000 rackmount boxes.

I am still skeptical that we spend twice the power cooling the datacenter than we sprnd running the servers.

And it's Kentucky, not Kansas.
Re:Also I wonder by evilviper · 2003-08-23 17:26 · Score: 1

Tripple that to allow for cooling (It takes approx 2 watts of power to remove the heat generated by 1 watt of power usage)

Ahem... I have sever computers running constantly, and can get away with NO extraa energy for cooling. You see, those fans in the back of your power supply work quite well, and if they are attached to ducting that will direct the heat to some harmless location, you can get away with each computer only outputting a tiny fraction of the heat into the room.

Incidentally, they would probably be better off sealing that room off from the rest of the building, and removing the roof entirely, than trying to use air conditioners to cool everything.

They could even (instead of ducting hot air out) duct cool air in to the system, which is a practice I've seen successfully done where computers need to operate in open-air environments, here in the desert.

p.s. more wire does not meed more juice,

That's just simply false. The laws of physics says the more wiring, the more power needed. The more correct answer to the question would be that the electricity on those wires is quite tiny, so even multiplying that many times will only result in a nominal power draw, compared to the enormous ammounts of power used by processors, hard drives, etc.

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Re:Also I wonder by evilviper · 2003-08-23 17:58 · Score: 1

Well, Athlons typically want 350W.

Says who? My Athlon XP 2000+ system, with 7200RPM hard drive, is drawing just over 100watts regularly.

I've recently tasked out the problem of power distribution, and the solution I came up with was gutted DC AVR'd UPS's (with the DC->AC component gutted) that directly connect to the motherboard; bypassing the switching powersupply and any external drives. Given enough external cooling, you could even resort to passive heat-sinks (albeit large/copper/expensive)

Holy crap man, I'm not even sure where to star poking holes in your idea.

First of all, UPSes don't provide all the voltages you would want, meaning you'd need an additional power supply to step-down the voltages. That isn't exactly going to be cheap. The reason you can get power supplies for $15, is economies of scale, and you aren't going to get that with your custom DC-DC voltage adapter.

Also, you are going to need incredible accuracy. With 120v building wiring, the voltage normally fluxuates sometimes as much as 5 volts, even comming from a UPS (resistance and load has serious effects). With high voltages, that's no big deal, but with your DC supplying multiple system, direct to the mobo, you obviously can't have 5 volt fluxuations on your 1v, or 3v lines, even just for a millisecond.

As for passive cooling, systems just aren't designed for it. The only way it would work in a standard case, is if you cut the entire top of the case (and power supply) off. I know this, because I have found myself with passive-cooled systems many times when fans have failed. Larger heat-sinks may help the processor, but not enough, and the power supply is *certain* to overheat unless it is specifically designed to be fanless (which are big-dollar items).

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Re:Also I wonder by DoctorRad · 2003-08-24 00:57 · Score: 1

Yeah, I wondered a while back whether if you factored electricity costs into TCO, whether it might be cheaper in the long run to use a larger number of lower-powered but higher MFLOPS/Watt processors.
Anyone done the math?
Dr. Matt...

--

Save the Bottom Line

To those who might not know... by qrash · 2003-08-23 02:25 · Score: 2, Informative

gigaflop

As a measure of computer speed, a gigaflop is a billion floating-point operations per second (FLOPS).

--
you may find the Higgs in this signature.

Re:To those who might not know... by fuzzix · 2003-08-23 03:15 · Score: -1, Offtopic

bellyflop Swimming pool smackdown that will wind the fuck out of you. Measured in splash operations per second (SPLOPS)
Re:To those who might not know... by ant_slayer · 2003-08-23 04:30 · Score: 2, Informative

If you're going to try to be informative, at least be accurate. There's no such thing as a "gigaflop". That would mean "Billions of Floating point Operations Per..." without the unit of time.

It's a gigaflops (singular). The 's' is very important. It's how we know how long it takes to perform a billion floating point operations.

It's like when people say "I had my engine up to 6000 rpms". What's an rpms? Is it a plural rpm? If so, what is pluralized? The acronym expansion yields "revolutions per minute", so would it be "revolutionses per minute", "revolutions per minutes", or "revolutions per minutes"s? None of 'em make sense. Technically, anything that revolves revolves at 6000 revolutions per some number of minutes... oi.

The Earth rotates at 6000rpms... if the unit of time is in blocks of 8640000 minutes... This type of confusion is why the time unit in our units of speed is usually unity.

-Josh O-
Re:To those who might not know... by Lozzer · 2003-08-23 09:28 · Score: 1

1 rpm = 1 revolution per minute

2 rpms = 2 revolutions per minute

--
Special Relativity: The person in the other queue thinks yours is moving faster.
Re:To those who might not know... by Anonymous Coward · 2003-08-23 10:03 · Score: 1, Funny

computer

A beige box with wires coming out of it.

Let's not get too excited.... by CGP314 · 2003-08-23 02:25 · Score: 5, Funny

Supercomputer Breaks the $100/GFLOPS Barrier

Not after you factor in the SCO license fees.

Re:Let's not get too excited.... by qrash · 2003-08-23 02:33 · Score: 0

That is $100 minus $699/GFLOP

--
you may find the Higgs in this signature.
Re:Let's not get too excited.... by CGP314 · 2003-08-23 02:39 · Score: 1

Hey!

That's not funny, that's informative

-Darl McBride
Re:Let's not get too excited.... by McBride,+Darl · 2003-08-23 03:00 · Score: 0

The parent is yet another poser, and is not in fact Darl McBride. My lawyers inform me that a cease and desist letter regarding this matter will be sent via facsimile to OSDN later today.

--
Darl McBride
Chief Executive Officer
Caldera International, Inc.
Re:Let's not get too excited.... by Alien+Being · 2003-08-23 03:16 · Score: 1

Imagine a beowulf cluster of subpoenas from Some Cokamamey Outfit.
Re:Let's not get too excited.... by Anonymous Coward · 2003-08-23 05:27 · Score: 0

When I saw this post, I imagined a scene from the movie Better Off Dead with SCO as the kid on the bike:

2 dollars! I want my two dollars!

Hey, works for me by DarkSarin · 2003-08-23 02:25 · Score: 1, Redundant

I personally love to see this kind of stuff. As a kentucky native, anything related to my home turf gets extra kudos. On a more technical note, I just want to know why I can't have one of these?

--
"We don't know what we are doing, but we are doing it very carefully,..." Wherry, R.J. Personnel Psychology (1995)

Re:Hey, works for me by Drakin · 2003-08-23 03:09 · Score: 1

Your home turf? The story wasn't about kentucky bluegrass...
Re:Hey, works for me by DarkSarin · 2003-08-25 14:37 · Score: 1

I am actually from Kentucky (eg UNIVERSITY OF KENTUCKY), and I grew up around the University, so yes this is my home turf. The real question is why am I responding to you??

--
"We don't know what we are doing, but we are doing it very carefully,..." Wherry, R.J. Personnel Psychology (1995)
Re:Hey, works for me by Drakin · 2003-08-25 15:05 · Score: 1

Beacuase you've got zero sense of humour? It's a joke... a play on words...

fuck you man, i made it by Magic+Thread · 2003-08-23 02:26 · Score: -1, Offtopic

Because you cannot afford the electric bill. by HanzoSan · 2003-08-23 02:27 · Score: 1

Not to mention I'm 100% sure you use Windows, and cant even take advantage of the power of that.

How would you use this super computer?

--
If you use Linux, please help development of Autopac

Re:Because you cannot afford the electric bill. by Kurt+Russell · 2003-08-23 03:00 · Score: 1

How would you use this super computer?

I like this guy's idea.
play chess with a Natalie Pr0tman hologram.
yes sir I do..
Re:Because you cannot afford the electric bill. by thynk · 2003-08-23 19:12 · Score: 1

How would you use this super computer?

The same way I use every other computer at home. pr0n, lots and lots of pr0n. Might make a nice Counter Strike server too.

It's a sure sign that you're a geek when people ask what the first thing you'd do if you won the lotto is, and you answer "I'd build my own Beowulf cluster" is the first thing that comes to mind.

--

Good judgment comes from experience, and a lot of that comes from bad judgment.

It depends on which kernel they used by Allah · 2003-08-23 02:27 · Score: 1

the older kernels, they are SCOless......

Re:It depends on which kernel they used by negacao · 2003-08-23 02:39 · Score: 1

*shrug* so are the newer kernels.

don't worry, SCO will eventually decide that the entire 2.x series is infriging on their quantam chess playing application.
Re:It depends on which kernel they used by CaptBubba · 2003-08-23 03:06 · Score: 1

According to the second link, they are using 2.4.21.

It's a university project by Anonymous Coward · 2003-08-23 02:29 · Score: 3, Funny

Remember, everyone, this was a university project. *BSD was also a university project originally, and now *BSD is dying. So obviously university projects are not of very high quality.

Re:It's a university project by krahd · 2003-08-23 02:37 · Score: 1

yeah sure, I've started and later managed/directed a project which has failed completelly. So obviusly solo projects are not of very high quality. Linus: face that and leave that linux thing.

--krahd

--
mod me up scottie!
Re:It's a university project by eric76 · 2003-08-23 02:42 · Score: 1

Just wait til we have OpenBSD on an XBox.
Re:It's a university project by Anonymous Coward · 2003-08-23 03:05 · Score: 0

obviously you didn't understand your parent post. hence you don't understand anything. hence you should have your brain removed. thank you very much for complying.
Re:It's a university project by Anonymous Coward · 2003-08-23 08:44 · Score: 0

The reason that *BSD is dying is not because it was a university project. It's dying because of the problematic personalities of many of the key players.
Re:It's a university project by Anonymous Coward · 2003-08-23 09:50 · Score: 0

I talked with your Dad about this and _you_ were 'a university project' too. He mentioned something about a party where ... err, ummm ... it would be better if you asked him yourself sometime when your Mom is not around.
Re:It's a university project by Anonymous Coward · 2003-08-23 14:13 · Score: 0

Netscape ... is dead. It was the Undergraduate Students solution.

Internet Explorer ... is Alive. It was, Mosaic, the Graduate Students solution.

Yes, you had to be there to believe it.

X

Asymmetric Sparse Flat Neighborhood Network by FreeLinux · 2003-08-23 02:29 · Score: 5, Interesting

Obviously, I don't get it. This doesn't look any different than redundant backbones or what is frequently done with VLANs. Multiple paths between hosts is what I see. How is this "new"?

Re:Asymmetric Sparse Flat Neighborhood Network by flymolo · 2003-08-23 02:38 · Score: 5, Informative

Due to "creative" (computed) wiring, if all switchs are functioning, no node is more than one hop from each other node. This requires a routing table written for each pc. It could be used for redunancy, but it is being used to minimize latency, and collisions, which are both killers in clusters.

--
"Sometimes it's hard to tell the dancer from the dance." --Corwin Of Amber in CoC
Re:Asymmetric Sparse Flat Neighborhood Network by bengoerz · 2003-08-23 02:41 · Score: 1

Actually, you're exactly wrong in that there are NO multiple paths between any of the hosts -- just one path between each. So what has already been responded is right -- it's all about single-hop low latency results.
Re:Asymmetric Sparse Flat Neighborhood Network by FreeLinux · 2003-08-23 02:48 · Score: 2, Interesting

no node is more than one hop from each other node. This requires a routing table written for each pc.

Admittedly, I understand that no node is more than one hop away. But, how is this different than all nodes plugged into a large switch like a Cisco 6500 or a Nortel Passport 8600? These switches can have ~128 ports and can switch 256Gbps aggregate throughput at wire speed. Add another switch and then add a second NIC to each host and you increase the capacity even further. Additionally, this does not require special routing tables or software on the hosts.

The technique that was used seems to be more of a mental exercise in making spaghetti, I don't see it reducing latency or increasing performance beyond the currently used techniques.
Re:Asymmetric Sparse Flat Neighborhood Network by Arker · 2003-08-23 02:55 · Score: 1

But, how is this different than all nodes plugged into a large switch like a Cisco 6500 or a Nortel Passport 8600?

It's cheaper.

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
Re:Asymmetric Sparse Flat Neighborhood Network by Rich+Dougherty · 2003-08-23 02:57 · Score: 3, Informative
Here's a quote from the site:
Does The World Need Yet Another Network Topology?
One would think (well, we did ;-) that the latest round of Gb/s network hardware would have made the design of a high-bandwidth cluster network a trivial exercise. However, that isn't the case when the prices are considered:
- When we invented FNNs in 2000, the cheapest of the Gb/s NICs available were PCI Ethernet cards priced under $300 each; now they are $50-$100. Prices have continued to drop. Prices on custom high-performance NICs (e.g., Myrinet) start at close to $1000 and have not been going down.
- In late 2002, 48-port 100Mb/s Fast Ethernet switches have dropped to less than $25/port. Gigabit Ethernet switches are starting to follow the same trend, with $100/port pricing in sight for switches up to about 48 ports. Wider switches with the needed performance are unlikely to become cheap in the near future. Thus, it would be necessary to build a heirarchical switch fabric using multiple layers of switches, yielding higher cost, higher latency, and significantly lower bisection bandwidth (unless you use a "fat tree" or other scheme, which adds still more expense -- especially because cheap layer 2 Ethernet switches don't support those topologies).
In summary, the cost of the "obvious" Gb/s network for KLAT2's 66 single-processor nodes was OVER 30 TIMES the cost of the network we built for KLAT2. In fact, to match KLAT2's bisection bandwidth, a network built using Gb/s hardware would have cost even more. Gigabit Ethernet is getting cheaper, but obvious topologies just are not competitive with FNN performance. So, if you've got tons of money that you have to spend immediately, you can impress your friends by buying expensive custom network hardware that can use an obvious topology and still be competitive with FNN performance. Otherwise, read on.... ;-)
Re:Asymmetric Sparse Flat Neighborhood Network by guile*fr · 2003-08-23 03:01 · Score: 1

I suppose that kind of switches would cost alone more than 39,000 $
Re:Asymmetric Sparse Flat Neighborhood Network by Anonymous Coward · 2003-08-23 03:03 · Score: 0

It's cheaper.
Bingo!
Except, this is like the clusters were I work. The parts are cheap, but the cost to design, impliment, debug, and maintain them is through the roof. To bad it's on a different budget, or they wouldn't be so wasteful.
Re:Asymmetric Sparse Flat Neighborhood Network by cperciva · 2003-08-23 03:36 · Score: 1

Traditionally, people have tried to keep their routing tables small. When you're routing in hardware, the larger your routing table, the slower (or more expensive) your routing hardware is. As a result, you want to have single routes which apply to entire groups of hosts (eg, "packets for nodes 0-127 go through port 0, packets for nodes 128-255 go through port 1").

Because the routing is being done in software instead, the cost driver is dramatically reduced; consequently, it becomes cost-effective to have a routing table with an entry for each node.

Note that even in software, this approach doesn't really scale: If each node has a table entry for each other node, the memory cost is quadratic. This isn't a problem for a few thousand nodes, but this sort of routing would never work for the internet.

--
Tarsnap: Online backups for the truly paranoid
Re:Asymmetric Sparse Flat Neighborhood Network by Anonymous Coward · 2003-08-23 04:04 · Score: 0

That's all well and good.

It's STILL not a new network topology, and it hasn't been since 2000.
Re:Asymmetric Sparse Flat Neighborhood Network by sjames · 2003-08-23 05:37 · Score: 1

The technique that was used seems to be more of a mental exercise in making spaghetti, I don't see it reducing latency or increasing performance beyond the currently used techniques.

It significantly reduces cost. In wire speed switches (FastE or GigE) there will typically be a sweet spot for price/performance. Beyond that point, switch prices jump into the stratosphere.

For larger clusters, there simply aren't any switches big enough at any price (just try to get a 256 port GigE wire speed switch for example). In that case or if you're not made of money, you must either live with sub GigE bandwidth and higher latency, or get creative (Such as ASFNN).
Re:Asymmetric Sparse Flat Neighborhood Network by jpc · 2003-08-23 06:14 · Score: 1

(they have 64 machines, not 128, so I have done the numbers with this).

you can increase performance. rather than 1 Gb port into a very expensive 64 port switch, to give you a maximum of 128Gb bandwidth (bidirectional 64x1Gb), you can (if you use the calculator) stick 4 Gb ports in each machine, buy 11 cheapo Dell 24 port gigabit switches (about $3k each), have 1 switch latency, and have 4 times the total non blocking bandwidth available. And the switches will still cost you less than 1 64 port gig switch. Not a *huge* amount less though.

Of course they used 100Mb switches which are virtually free, which with 4 NICs gives you more bandwidth than a tree of gig switches but less than the very expensive 64 port solution.
Re:Asymmetric Sparse Flat Neighborhood Network by kasperd · 2003-08-23 10:32 · Score: 1

Because the routing is being done in software instead, the cost driver is dramatically reduced; consequently, it becomes cost-effective to have a routing table with an entry for each node.

I was actually wondering how well Linux would handle this. The obvious algorithm to find the correct entry in the routing algorithm is linear in the number of entries. That doesn't sound like efficient to me, but it might be that 100 entires is still so small a number, that it doesn't matter. However this particular case can be done far more efficient than the general purpose algorithm. Because the number of possible targets is rather small, it is feasible to tabelize the IP->interface choice, it would take less than 1KB of RAM. Of course that algorithm would never work if you were to route packets to any destination on the internet. Did they have to modify the routing in the kernel? And how exactly is it done?

--

Do you care about the security of your wireless mouse?
Re:Asymmetric Sparse Flat Neighborhood Network by jpc · 2003-08-23 23:45 · Score: 1

Just because there are not multiple routes doesnt mean the overall bandwidth is not higher, becasue as you have multiple nics, you can send to multiple hosts at the same time, so the total bandwidth is higher. Obviously it does depend on your communication pattern how much you can benefit from this.
Re:Asymmetric Sparse Flat Neighborhood Network by bengoerz · 2003-08-24 02:43 · Score: 1

When I toured their previous cluster, these guys completely touted the benefits of low latency. They didn't really care about the bandwidth -- at least it wasn't one of their major concerns. In fact, one of the guys (Tim) had done major work on a Parallel port connection device that he said had extremely low latency. However, it was limited in how many nodes it could handle, and so I believe that they went with ethernet to take more nodes. Otherwise, this thing might have been connected by the old parallel ports!
Re:Asymmetric Sparse Flat Neighborhood Network by Anonymous Coward · 2003-08-24 06:00 · Score: 0

It could be used for redunancy, but it is being used to minimize latency, and collisions, which are both killers in clusters.

Right. I guess the problem is that it will completely ruin the stability of the cluster. Any time a single machine goes down, parallel jobs die and the machine won't work again until you have replaced the faulty node with a spare one.

With 128 nodes this will likely happen about once a month, which essentially makes this useless for production supercomputers...

Students as Slave Labor by gremlin_591002 · 2003-08-23 02:29 · Score: 5, Funny

Ponders while there are not University students pictures in the National Geographic Article on Slavery....

this is nice by the_2nd_coming · 2003-08-23 02:30 · Score: 2, Interesting

but super computers as in giant iron are becoming more specialized and as such would woop the pants off a Beowulf cluster when competing in the specialty.

of course, if you just need a lot of general purpose super computing, it is obvious that you cannot compete with this.

--

I am the Alpha and the Omega-3

Re:this is nice by stdarg · 2003-08-23 12:09 · Score: 2, Insightful

Wait until computers start shipping with a few FPGA units. Then you can flash a new image onto the FPGA's for each specialized application you use the cluster for.

Playstation2 at 5.5GFLOPS costs only $199 $40/GFL by gorim · 2003-08-23 02:31 · Score: 4, Insightful

And it was introduced to consumers just a couple years
ago. Sorry, the AMD beowulf cluster at $100/GFLOP just
isn't that impressive.

The burning question by Timesprout · 2003-08-23 02:32 · Score: 2, Funny

though is how many mp3's are these students sharing on this monster ?

--
Do not try to read the dupe, thats impossible. Instead, only try to realize the truth
What truth?
There is no dupe

Re:The burning question by HanzoSan · 2003-08-23 02:43 · Score: 3, Insightful

People dont share mp3s anymore, if they do the FBI, NSA, Secret Service, CIA, and Homeland Security Dep will swarm them and put them in the bay.

I mean I wish we could crack down like this on organized crime, or on domestic terrorists, I'm surprised we are so aggressive at arresting teenagers who download music, but the KKK and Neo Nazis can collect a million guns and spread their crazy hate speech and its protected by freedom of speech.

I'd think that hate speech does more harm than copyright infringement.

--
If you use Linux, please help development of Autopac
Re:The burning question by zeroclip · 2003-08-23 03:14 · Score: 1

Problay none since the whole cluster dosent have any storage space. Unless you include the 512mb ddr ram on each node :P
Re:The burning question by Anonymous Coward · 2003-08-23 03:31 · Score: 0

Fuck off, faggot! And take that other faggot, Dean, with you.

Stop this super-cpu insanity! by Anonymous Coward · 2003-08-23 02:32 · Score: -1, Troll

It saddens me that such renowned researchers have to resort to fruitless projects such as these to fulfill their academic quota.

Their time and efforts would be better server if they applies their knowledge to more tangible projects that would have a direct benefit to society. It is imperative that we encourage the new generation of scientist to pursue these lofty goals. Only then can we wrest ourselves away from the stranglehold of corporate-sponsored research that dominates the published literature.

Which is nice.

Re:Stop this super-cpu insanity! by Anonymous Coward · 2003-08-23 02:58 · Score: 0

Eliminating "the stranglehold of corporate sponsored research" will only serve to eliminate any meaningful applied research.
In my country there has been a lot of discussion about how "evil" it is that many university level theses that have been done in collaboration with corporations are declared classified (=won't be made public until a few years later).
I just would like to know what is the point in doing research if it does not benefit the society? If classified theses are no longer possible, the business will lose all interest in the academic research and the stereotype of a acadmic researcher will become reality. There should be a healthy balance between applied and basic research.

hot damn, they're case modders! by mrgreenfur · 2003-08-23 02:34 · Score: 2, Funny

each node has two side case fans! that's gotta be the most dedicated case modding job i've ever seen! 132 pc's with 2 fans! too bad they didn't put fan guards ... or interior lights.. or blue led's... but i guess all that junk about a supercomputer makes up for it...

Re:hot damn, they're case modders! by BubbleNOP · 2003-08-23 02:46 · Score: 1

That's not a mod, they just got this case.
Re:hot damn, they're case modders! by ptbarnett · 2003-08-23 04:43 · Score: 1

No, the case came that way. But, you will notice that they are stacked next to each other, blocking the side ports for all except the ones on the left end.
That's probably why they did this:
For example, each case came with two side fans, which we converted into a redundant stack venting out the back.
Re:hot damn, they're case modders! by commodoresloat · 2003-08-23 08:44 · Score: 1

yeah man where are all the blinkenlights?

Thats not new, but no its not slave labor. by HanzoSan · 2003-08-23 02:34 · Score: 1

The government does give financial aid to some of them, along with scholarships and other forms of payment, so they get to live for free on campus with government money, its kinda like the military deal, you work hard, you get an education, and the government supports you.

In this situation I'm sure these kids, all of them, will win all kinds of scholarships.
(payment)

--
If you use Linux, please help development of Autopac

YOU FAIL IT! by handybundler · 2003-08-23 02:34 · Score: -1

in light of your failure, your homeowrk assignement for the weekend will be to imagine constructing a beowulf cluster of these and using it play chess with a Natalie Pr0tman hologram.

Quite simply, YOU HAVE FAILED.

--

a/s/l here. Sorry, adding domain tags to your s

MODS ON CRACK COCAINE! by Anonymous Coward · 2003-08-23 02:36 · Score: -1, Offtopic

This is a terrible system, this moderation done by simians.

Simpsons by CGP314 · 2003-08-23 02:37 · Score: 0, Troll

Looks like Frink was a bit off:

Well, sure, the Frinkiac-7 looks impressive, don't touch it! But I predict that within 100 years computers will be twice as powerful, 10,000 times larger, and so expensive that only the five richest kings of Europe will own them.

Why do you always call it slave labor? Its not. by HanzoSan · 2003-08-23 02:38 · Score: 0, Flamebait

I think working at mc donalds is slave labor, you cannot pay your rent, you cannot buy your food, you cannot survive, thats slave labor. I dont think this is slave labor, students live better than a person who works at mc donalds. Students do get support from the government via financial aid, they do get scholarships, they get stipends, fellowships, and many many payment systems setup to help them.

Basically their full time job is "student" just like your full time job is whatever it is that you do. They do stuff like this and it pays their bills, they live off scholarships, you live off your paycheck, I'd rather be a student however than work at McDonalds, which is my option if I decide not to go along with the "slave labor camp" .

So you can be a student slave, or a corperate slave, which one do you prefer?

--
If you use Linux, please help development of Autopac

Re:Why do you always call it slave labor? Its not. by Toby+Studabaker · 2003-08-23 02:49 · Score: 1, Interesting

So you can be a student slave, or a corperate slave, which one do you prefer?
Where I live you can also sit on your ass on government welfare and do nothing.
Which, I think, is perfectly OK. I'll happily pay taxes to keep these people housed and above the poverty line than have them begging on the streets and mugging people for a living. People who don't want to work/won't work are, after all, a minority and will never bring a society financially to its knees.
And before the libertarians jump in and start spouting crap about how they should not be made to pay for these freeloaders, let me remind you that you are free to leave the society. You can take your business and live outside it, all tax free, but you'll come crying back when you realize that you'd have to fund your own law enforcement, health care and rescue services.
Re:Why do you always call it slave labor? Its not. by ctxspy · 2003-08-23 03:19 · Score: 1

oh ok, so then you are in favor of accepting the status quo in 100% of cases, never wanting to change anything?...

"if you dont like it leave!"?... If it weren't for people who thought the exact opposite way, America wouldn't exist in the 1st place for you to spout off your point of view.
Re:Why do you always call it slave labor? Its not. by HanzoSan · 2003-08-23 03:23 · Score: 1

"Where I live you can also sit on your ass on government welfare and do nothing."

Havent you heard of welfare reform yet? People on welfare must work 40 hours a week at Mc Donalds.

--
If you use Linux, please help development of Autopac
Re:Why do you always call it slave labor? Its not. by Anonymous Coward · 2003-08-23 03:31 · Score: 0

The single most humane post I have ever seen on slashdot.

Thankyou for knowing what living in a society means and not falling for the social darwinism so many seem to support.
Re:Why do you always call it slave labor? Its not. by Toby+Studabaker · 2003-08-23 03:48 · Score: 1

status quo in 100% of cases, never wanting to change anything?
I don't know where you got this "change anything bit", but if change means kicking people out to the streets because they can't/won't work the hours/jobs that are available, then no I don't want to change anything.
if you dont like it leave!
No, I simply stated a fact: if you don't like paying taxes in order to take care of the troublesome/sick minority you can always leave - the minority can't. You won't starve, they will.
Re:Why do you always call it slave labor? Its not. by jd · 2003-08-23 04:12 · Score: 0, Offtopic

Personally, I don't like the idea of anyone being a slave to anything.
I grew up in England at a time when the Government still payed 100% of the tuition fees AND a very sizable grant for living expenses (I got something like $8,000 per year).
Compare that to the US, where students need to excell at sports, or wage-slave themselves between lectures to keep going.
Slavery was "abolished" over 200 years ago. It is time to finish it completely. Students, when they graduate, will often earn better jobs and therefore pay higher taxes. Charging them for learning, then for living, and then charging them yet again for student loans, is to tax them four times over for learning.
This is why only the rich can learn in the US. Very very few poor people have degrees, unless they are on a sponsorship trip, in which case they are likely to be spending too much time on the game field to be really learning a damn thing. Sure, it gets them out of the ghetto, but it still ain't learning. Don't mistake the two.
I regard the Thatcher years as the latter-day "Dark Ages" of Britain, but I think she got a few things right. She wanted to increase the number of people getting into University, and aimed to make it practical for everyone to make the choice of going, if they wished.
These days, I believe the estimates place the number of people who have the option of going to University in the UK at around 60%. We're not talking vocational colleges, or unaccredited e-mail spamiversities. We're talking real, established, often hundreds of years old, centers of learning with a proven track record.
The US doesn't even come close. Many kids can't even get through high school, because their parents can't afford it. Public schools in the US are a joke - assuming you survive the endemic violence in them.
There is no possibility of a "reneissance" in the US, as it is utterly counter to the existing system. Power is held by the rich, the rats are kept by the poor. The US, far from being classless, is an excellent example of a fiefdom.
Yes, I resent that. I resent the fact that students are cheap "slave labor". I resent the fact that those who do the most are paid the least and get the least credit. I resent the fact that the "land of opportunity" exists only in the minds and pockets of the rich elite.
Why all this resentment? Because I come from a land which had learned the errors of this way of life. It learned the hard way, by becoming virtually bankrupt, by destroying virtually the entire of its own society, that what happens to the poor in society will eventually happen to the whole of society. You can run, but you can't hide from the monsters you create.
So why am I in the US at all? Because the US is still better off than the UK. The UK learned, but too little and too late. It is still decaying, as evidenced by the expulsions of tourists for not knowing who Yoko Ono was. The spy cameras, the paranoia - these things are proof positive that the UK is still in a state of decay. The better education will help, eventually, but it'll take time. The UK won't recover in my lifetime, though.
The US, on the other hand, is not yet in that same state of total ruin. It's a lot bigger, and a lot richer, which helps stave off these problems. The continuing evidence of yet more accounting fraud, the fragility of the power networks and the total absence of any serious mass transit for most of the country means that further decay is inevitable, and may be irreversible by this point. Nonetheless, the US is still a country you can live in. The UK really is not.
So if I dislike the US this badly, why stay here at all? What are the alternatives? New Zealand has a low crime rate, true, but getting in is next to impossible and there's no industry there, to speak of. Australia is set firmly on the same course of self-destruction as the UK and US. Most of Europe is in bad shape

--
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Re:Why do you always call it slave labor? Its not. by Anonymous Coward · 2003-08-23 05:07 · Score: 0

It's a myth that the US federal government helps students out. The only ones who get financial aid are the ones whos parents make less than something like $20,000 a year or the students who are great at sports (we gotta have them sports players bringin in money for them universities, now don't we? Sure glad US has its priorities straight.)

Of course you will read somewhere (collegeboard.org or some such) that anyone can go to college and it's not just for the "rich." The only thing they forget to tell you is that you will be paying $40,000 or more in student loans when you get out. I'd say this is 80% of the students *right here*.

Getting a scholarship is simply a matter of winning a popularity contest. Ever notice how the ass-kissing students always get the better grades? One thing leads to another and they eventually have a scholarship. This is probably 2% of the student population.

Getting grants from filing a FASFA form requires having parents which *do* work at McDonalds. Or perhaps, having no parents at all. Grants are good for a total of roughly $1,700. Tuition + books + parking permit is roughly $2,500. Thank you US government! You will also probably qualify for government loans (subsidized or not). This means you get roughly another $1,700 which you pay back to the government. Your interest rate will be something like 4.5% Thanks, but I can get $5,000 or more on 0% interest credit cards and simply transfer the balance when need be. If you decided to live on-campus you have to figure that the cost of living is going to be roughly $4,000-5,000 a year. Even more if you live off-campus.

The government is *never* helping students out. You want to know who helps students? Corporations if anything. They give scholarships and they given internships and work-study type programs. You can make *actual* money doing these. Many times it will be a minimum of $10 an hour too. Wonder how much you get stuffing envelopes for Uncle Sam...
Re:Why do you always call it slave labor? Its not. by nedjacket · 2003-08-23 05:13 · Score: 1

Uhh... This maybe flame-bait, but I can't pass this up.
What planet are you from, anyway?
This is why only the rich can learn in the US. Very very few poor people have degrees ...
I will grant you, there are stories of poor people who cannot get an education because they live in a cycle of violence or poverty, but this is hardly the standard. Rather than "The system won't let poor people can't get an education, therefore they're always poor", I submit that "People who didn't go for an education stay poor."
When I was getting ready to attend college, (1979) my father had been diagnosed with Alzheimers and was no longer working. My mother worked as a bookkeeper at a local radio station. I went to a 1st rate, 4 year school and got a good education. I borrowed money, I saved money working during high school and I worked summers to make it through. My situation is hardly special or unique. THIS is the standard... and there is nothing wrong with working for something.
I'm sure someone will pop in with "that was then, this is now". Well, now my state proves the HOPE scholarship. If you maintain an A or B average, your tuition is paid for. This is paid for by a state-wide lottery. "Lottery n. - a tax on the mathmatically challenged"
There are plenty of things that our society can improve on, but hopping in here with doom & gloom over why the government doesn't just pay everbody to go to college shows a tremendous lack of thought on your part. Socialism may sound great, but it rarely yields vibrant, properous societies where people actually have a chance to better themselves in life.
Re:Why do you always call it slave labor? Its not. by cymen · 2003-08-23 05:40 · Score: 1

Compare that to the US, where students need to excell at sports, or wage-slave themselves between lectures to keep going.

Stafford Loan Maximums (independent student):

1st year: $6625
2nd year: $7500
3rd year: $10500
4th year: $10500
5th year: $10500

For dependent students the amount is almost halved but any student can become indepent when they are 23 years old or demonstrate that they truely are no longer supported by their parents/guardians (don't know too much about this though). This is only Stafford loans too, there are plenty of other sources for loans.

Now if you had said loan-slaves maybe I'd be slightly more inclined to agree...
Re:Why do you always call it slave labor? Its not. by Anonymous Coward · 2003-08-23 05:46 · Score: 0

I saved money working during high school and I worked summers to make it through.
But wouldn't it have been nice if the society would have helped you with this and you didn't have to bother about paying back your loans?
Re:Why do you always call it slave labor? Its not. by ctxspy · 2003-08-23 05:48 · Score: 1

The key is to strike a balance between social help and helping oneself.

As far as who can and can't leave, ... umm... these minorities were probably in a WORSE situation where they came from, and you guessed it, they left... they came here.. where people are willing to pay them to sit around and do nothing.

A much more fruitful application of social welfare would be something that has recently started to gain momentum -- workfare.. in other words, you still have to work for a living, but you get a little assistance in bringing you above the 'poverty line'.

The sentiment of the current welfare system is nice, however it was implemented by a bunch of people who would rather look nice for the camera than think about the implications of their actions. The current plan is rife with abuse, and i think that is common knowledge. It encourages people to be lazy, sit on their ass (if you work, you get less / no assistance?), it also encourages having more kids (more kids = more money), however it doesn't require that you spend this money on raising your kids, or planning for your future, etc. All in all, sounds like a poorly implemented system.

Libertarians are a little out-there with their everyone-for-themselves (almost republican?) attitude, however the liberals are just as bad with their bleeding hart b.s. that you should be guilted into paying for other's lack of motivation.

Also, i'd like for you to explain why leaving is a better alternative to voicing your opinion regarding the use of your tax dollars.
Re:Why do you always call it slave labor? Its not. by jd · 2003-08-23 07:06 · Score: 1

You worked for your education. You worked again to pay for it. If you had student loans at the end, you worked a third time to pay those loans off. Then you worked a fourth time to pay the taxes which pay for all this infrastructure.
Oh, I agree your situation isn't abnormal. What I find abnormal is paying 400% over the odds (plus interest) and considering it a good deal.
I'm not talking about a socialist society, where everything is covered by "the state", but rather of any system (regardless of form or politics) where you pay for what you get, and you get what you pay for.
Hey, people are going to want a profit margin. Without profit, not many people have motive. Most shops are happy with a 50% mark-up. Specialist dealers (eg: technology) often work on a 100% mark-up, because they've less volume.
But why are you paying a 300% mark-up for a high-volume, relatively low-cost enterprise? That's the bit I don't get. And if you have more people who are educated, the cost of that enterprise per person goes down. (It costs about the same for a salaried lecturer to teach one person, a hundred or a thousand. As lectures aren't interactive, the quality isn't affected that much, either.)
In other words, three times the kids in University, with only a 100% mark-up, would give the Universities about the same profit, cost them about the same to do, resulting in a massive jump in skilled workers, resulting in greater revenue generation, producing more taxes, resulting in a wealthier nation.
Menial laborers - the apple-pickers of the world - generate next to nothing for themselves or their country, and produce nothing for their company that a machine couldn't do better, quicker and cheaper.
Abslutely nobody needs to exist in this class. There is nobody so uneducatable and so incapable that they can never perform any operation requiring some degree of skill, intelligence or other human quality. Even many of the "untreatably mentally ill" who are in permanent care are, in fact, perfectly treatable and perfectly capable. They cost us money in the long-term, because we don't invest for the long-term.
Socialism tends to do badly, because people don't tend to plan long-term. Capitalism does just as badly for the same reason. The bottom line is that the system is really irrelevent. It's what you do with it that counts.
I believe that stupidity and ignorance are the two greatest drains on the pockets of everyone, whether on the level of an individual or as part of an organization, or even as part of a country. You can't do much about stupidity, but you sure as hell can eliminate a lot of ignorance.
To tie this in with the supercomputer effort, I believe that this kind of research is extremely valuable - if it is designed to educate the maximum number of people practical. When research is done to give one or two senior staff members an extra citation (citations are often used to determine the market value of academics), then it has profited but a single person.
Constructing a whole new design of networking topology could be used to educate a whole new generation of engineers. Remember, even Seymour Cray was "just" an engineer, at one point. In many ways, he still was, when he died. The difference was, he knew enough to be really, really good, and he knew how to apply that to become really, really good.
Network topologies are a major area of research, and are worth VERY big bucks. The blackout in the Northeast was a product of a very bad design of network topology. The Internet's backbone is also cruder than it has to be. The PhD grad student, whose thesis was mapping the topology of networks, became so dangerous to the Government that his work is likely to be classified.
Good designers are extremely rare, because very few people understand what it takes to make a good design. The students who assembled this supercomputer had the opportunity to

--
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Re:Why do you always call it slave labor? Its not. by EastCoastSurfer · 2003-08-23 07:58 · Score: 1

This is why only the rich can learn in the US.

I agree the rich have it easier, but that is true anywhere in the world. I don't think money is the reason a poor person will likely get less education than a rich person though. If you haven't noticed, scholarships, grants, and loans make it possible for anyone to go to school if they *want* too. The problem is that the drive and discipline to go to school must be instilled by the parents and the family.

To say that rich people are holding down the poor and keeping them from going to school is not true at all. In fact it is quite the opposite. Who do you think gives money to all these school funds for financially challenged people to go to school? I bet it is not poor people.

The problem with society as a whole today is not this division between rich, poor, or "working families." The problem is that more and more people feel they are owed something. That they are entitled to rewards without having to work for it. This idea clearly permeates itself thoughout U.S. culture. Consumer debt at all time highs, people graduating high school without knowing how to read/write, and not forcing people who come to this country to learn english are all symptons of the same problem. People wanting something for nothing.

I'm sure whole volumes could be written on what is causing this, for a lack of a better term, laziness. I'm also sure many people have their own ideas on how to fix it, but the cause and the fix would take us even farther off topic.
Re:Why do you always call it slave labor? Its not. by AceM2 · 2003-08-23 12:39 · Score: 1, Offtopic

Where I live you can also sit on your ass on government welfare and do nothing.

I'm glad you're proud of the fact that there are hundreds of thousands (conservative estimate) cheating the system and being allowed to be leeches when they could be working. Actually, since you're so glad about supporting deadbeats, I'll be more than happy to let you pay my taxes. I'll make a deal with you.. Just give me your checking account number and I'll let you pay for my groceries too.

I'm not just being an asshole either, I've honestly had to work with people collecting government funds, and I'll tell you.. For every hard luck case, single mother gets laid off and is really looking for a job, there are at least 3 deadbeats and 2 crackheads who have found ways of cheating the system to collect good money. The way our system is now, why NOT cheat it? I'm not totally against government support, I don't want to see that single mother get kicked out of her apartment, but I'm certainly not proud of the deadbeats in this country. Welfare and foodstamps aren't preventing muggers and beggers one bit, and I'd wager that most 'muggers' (>50%) are already on government support.

People who don't want to work/won't work are, after all, a minority and will never bring a society financially to its knees.

The minority you speak of is huge. They're sucking billions out of our economy, and the able bodied free loaders should be stopped. The working man and woman, no matter what they're doing as long as it's legal, should be proud of themselves. Just think of the programs which could be funded by spending the welfare/foodstamps/etc money on them instead of giving it away to people who do nothing to deserve it. Think of the extra tax money that would even be generated by the thousands of people that could be employed instead of freeloading.

And before the libertarians jump in and start spouting crap about how they should not be made to pay for these freeloaders, let me remind you that you are free to leave the society.

That's such a stupid comment.. It's like saying after I build and maintain a successful business I shouldn't be allowed in the building because people want to smoke inside and I don't want them to.

You can take your business and live outside it, all tax free, but you'll come crying back when you realize that you'd have to fund your own law enforcement, health care and rescue services.

1) I can protect myself if allowed. I don't want to have to use one, but I'd use a gun, baseball bat, kitchen knife, or pencil if I had to.
2) Uhm.. I already have to pay full price to go to the doctor, what's going to change? I'll still just walk into one and pay the price.
3) Okay you got me, if I fall into a well I'll just hope some good person walks by to save me.. It's a risk I'm willing to take

Don't buy into the garbage of if you want protection you have to have welfare for deadbeats, it just doesn't make sense. I'm tempted to ask if you even have a job and what it is..
Re:Why do you always call it slave labor? Its not. by thynk · 2003-08-23 18:49 · Score: 1

What I find abnormal is paying 400% over the odds (plus interest) and considering it a good deal.

Did they offer math at your school?? I only paid once *shrugs*

I'm not talking about a socialist society, where everything is covered by "the state", but rather of any system (regardless of form or politics) where you pay for what you get, and you get what you pay for.

Odd, I paid for an education, I got an education (not that it always shows). Yes, some of my taxes go to fund state schools, but come on - some of my taxes also go to support crack whores.

Your post kind of travels around a bunch of topics so I'm really not sure what point your trying to make.

--

Good judgment comes from experience, and a lot of that comes from bad judgment.
Re:Why do you always call it slave labor? Its not. by Natalie's+Hot+Grits · 2003-08-24 16:15 · Score: 1

By the time you are 23, you ought have already graduated.

That independance thing is BS too. You have to have been emancipated before you were 18, or your parents must be dead. Simply showing that your parents don't pay for your food and shelter do not suffice.

When you are a dependant student (under 23, 1 parent is alive, this is the definition as given in the FAFSA, it makes no difference of your parents income), your loans are halved. If you make 7,000 dollars at a part time job, they expect 5,000 of that to go to your schooling, and the other 2,000 to go for food/shelter? Give me a break.

I'm not bitching that this is a shitty system, but your simple view of student loans is utter bullshit. And its ignorant to try to spout that bullshit off like you know what you are talking about.

--
Two infinite things: your stupidity and mine. But I'm not sure about the latter. If my sig offends you, I'm sorry.
Re:Why do you always call it slave labor? Its not. by HanzoSan · 2003-08-24 21:15 · Score: 1

Thats a myth. First stereotype/myth is that only minorities are poor, when you have the trailorparks filled with millions of poor white americans in them who live off welfare.

Second myth is that people on welfare dont work, they do work, but when you work at MCDonalds you dont make enough money to raise children.

Abortion is legal but usually poor people dont have abortions because they need all the help they can get from their kids, often kids help them pay for their house.

No, I simply stated a fact: if you don't like paying taxes in order to take care of the troublesome/sick minority you can always leave - the minority can't. You won't starve, they will.

You have a point, Minorities arent usually allowed in the neighborhoods with low taxes, like the trailorparks, or the all white small towns, if minorities did flood these places then the same people who complain about social programs now will complain that too many minorities are moving into their communities.

--
If you use Linux, please help development of Autopac
Re:Why do you always call it slave labor? Its not. by HanzoSan · 2003-08-24 21:29 · Score: 1

The key is to strike a balance between social help and helping oneself.

Welfare is not free money, you cannot get it without being poor, and without working 40 hours, if you work at McDonalds yes you can get it, and why do people believe only minorities are on welfare when mostly white people are on it?

As far as who can and can't leave, ... umm... these minorities were probably in a WORSE situation where they came from, and you guessed it, they left... they came here.. where people are willing to pay them to sit around and do nothing.

What about minorities who have been here for generations, or minorities like native americans who were here first? You are assuming all minorities are textbook illegal immigrants from mexico, when you have plenty of minorities who have been here for generations who are poor due to segregation and all the racism that went on for the last few hundred years.

The only minority group which is successful is the Asians, and it just so happens that the Asians are the only minority group who werent completely taken over by Europeans and robbed of everything they own. So its more complicated than you make it sound.

A much more fruitful application of social welfare would be something that has recently started to gain momentum -- workfare.. in other words, you still have to work for a living, but you get a little assistance in bringing you above the 'poverty line'.

And thats what welfare has been for the last 10 years since Clinton(Not Bush) signed welfare reform.

"The sentiment of the current welfare system is nice, however it was implemented by a bunch of people who would rather look nice for the camera than think about the implications of their actions. The current plan is rife with abuse, and i think that is common knowledge. It encourages people to be lazy, sit on their ass (if you work, you get less / no assistance?), it also encourages having more kids (more kids = more money), however it doesn't require that you spend this money on raising your kids, or planning for your future, etc. All in all, sounds like a poorly implemented system."

Dude will you do your research and read the law Bush just passed. You must work 40 hours a week now to stay on welfare not 20.

Libertarians are a little out-there with their everyone-for-themselves (almost republican?) attitude, however the liberals are just as bad with their bleeding hart b.s. that you should be guilted into paying for other's lack of motivation.

The truth is, socialists and libertarians are the extremes, most people are a mix of both. We know neither of these plans work in the real world.

Also, name some people who live off welfare without working, I do know of a few people who do this, but its literally a few, like maybe 5-10% of all the people on welfare. Should we get rid of welfare just because a small percentage abuses it? Perhaps we should get rid of money in politics because it encourages politicians to be lazy and accept free stuff.

While we are at it lets get rid of their retirement and put it in the stock market, so we can laugh at them when they lose their life savings and retirement money in an Enron type situation. ha ha HA! Thats funny.

"Also, i'd like for you to explain why leaving is a better alternative to voicing your opinion regarding the use of your tax dollars."

What he is saying is while you can move into your tax free small town where rent is cheap and taxes are low, minorities if they were to flood your small town, your townfolk would get their shotguns and kick them out or they'd leave the town, its called white flight.

What the other guy is also saying is, its better to provide social programs than to build prisons. I agree wit him I'd rather pay higher taxes than have a high crime rate and build more prisons. I think while some poor people are lazy, most arent, in fact most work harder than you do, they just arent as educated.

Try working two jobs, I bet you have never done it.

--
If you use Linux, please help development of Autopac
Re:Why do you always call it slave labor? Its not. by ctxspy · 2003-08-25 00:20 · Score: 1

Good responses -- you have corrected me with regards to the welfare system. If things are the way you say, then i have no qualms with it. We obviously agree on how it should be done, the only difference was that i wasn't aware that it had been changed to operate in such a manner.

Asians aren't the only minority to do well.. How about indians (from india)?

I have a good understanding of why it is that certain minorities fare better than others (chiefly a cyclical problem passed down by the generations).

one thing you are incorrect about, is racism. Racism affected nearly EVERY new nationality of people that entered the country -- be it italians (damn plumbers), irish (heavy drinkers -- do you remember when they were specifically not given jobs for years?), arabs (towel heads), asians (more derogitory terms than i care to list), etc etc. Immigrants in general are often told to go back to their country of origin, without regard to the fact that all of us, including the 'native' americans are immigrants.

I dont deny that many immigrants work harder than I do..but, that is the advantage of having an education, and a specialized skill.

OH, and gues what? i'm an immigrant too... from an eastern european country ruled by the USSR til around 10years ago no less.

i'm not sure what your deal is personally attacking me with comments such as 'try working two jobs, i bet you have never done it.'... Have you? or are you too busy posting essays to slashdot. Oh, and i'm on my way to my 9 to 5.. while my mom is on her vacation back to hungary for the 1st time in 13 years -- and she's done the two job thing quite often. BUT, she did it so that i could get an education and not have to.

So maybe i should work THREE jobs just to make you happy? To prove that i am "hard working" enough to speak about this issue?.. sounds pretty stupid to me.

In closing, i'd appreciate if you re-read my post, then before replying with your knee-jerk reactions, try to comprehend what i've said.. Much of it is in-line with what you are attempting to get at, thus rendering most of your response redundant, and only leaving a few decent points and a few personal attacks.
Re:Why do you always call it slave labor? Its not. by HanzoSan · 2003-08-25 04:16 · Score: 1

one thing you are incorrect about, is racism. Racism affected nearly EVERY new nationality of people that entered the country --

YES BUT NOT EQUALLY be it italians (damn plumbers), irish (heavy drinkers -- do you remember when they were specifically not given jobs for years?),Yeah and some minorities are still not given jobs or equal salary in 2003 arabs (towel heads), asians (more derogitory terms than i care to list), etc etc. Immigrants in general are often told to go back to their country of origin, without regard to the fact that all of us, including the 'native' americans are immigrants.
Yes but some worse than others. Italians had the chance to someday become part of society, they suffered while they were immigrants, just like all the other white minorities you mentioned, the difference is, these minorities are all white, they had all their rights, they were not treated like "colored people" and segregated.

The reasons Asians are successful is because on average they are more educated than whites to such an extent that no one can overlook hiring them. Alot of asians are this successful because in their home country they picked the smartest kid in their whole family and told them "Go to America, get your education!" They were specifically taught from the beginning that they would not be treated fairly over here unless they were more educated than the average white male competitor.

When I talk about racism, I'm also talking about oppression, the only group of people besides colored minorities who have suffered the kind of oppression I speak of are the jews. Alot of jews were forced to change their name and pretend to be white in order to end that oppression. Now theres no way to know who is a jew, but alot of people on the college campuses are.

The reason italians moved past oppression is because they had other people to oppress, the mexican, the black, the chinese, the native american, in the end italians are still considered white. Italians do not have a very successful history in the USA, with the mafia and all of these setbacks, if there were not minorities here below their level, the prisons would be filled with italians right now instead of mexicans and blacks. Why isnt it filled with Italians even though italians run most of the organized crime orgs that the mexican or black perosn sells drugs for? Well thats because race does still matter, just look at the jails for proof.

I dont deny that many immigrants work harder than I do..but, that is the advantage of having an education, and a specialized skill.

What you ignore is the fact that a white male does not have to be as educated or as skilled as a minority to make the same amount of money. You ignore the fact that when the whites were facing racism, they could change their name, change their culture, and blend into society and become just a white person. Suddenly it all would end.

i'm not sure what your deal is personally attacking me with comments such as 'try working two jobs, i bet you have never done it.'... Have you? or are you too busy posting essays to slashdot. Oh, and i'm on my way to my 9 to 5.. while my mom is on her vacation back to hungary for the 1st time in 13 years -- and she's done the two job thing quite often. BUT, she did it so that i could get an education and not have to.

My point is, poor people work harder but always get called lazy while rich people work the least hard and dont get called lazy. If your mother worked hard, and your mother isnt lazy, and my mom worked hard and my mother isnt lazy, where do you get the idea that the poor are lazy?

Most people on welfare work hard, most of them have kids, yes some of them have kids on purpose to get money, but this is a small percentage. Alot of rich people do enron type things but should we move to socialism over it?

Also I'm not personally attacking you, the point I am making is, poor people and minorities get blamed for everything, while rich wh

--
If you use Linux, please help development of Autopac

You missed out on a great chance to be quiet! by Anonymous Coward · 2003-08-23 02:39 · Score: 0

Don't let that happen again. You will only harm yourself.

Thank you.

Imagine... by foobrain · 2003-08-23 02:40 · Score: 1, Redundant

... a beowulf cluster of these!

Re:Imagine... by Anonymous Coward · 2003-08-23 02:49 · Score: 0

my boot up'n your ass.

Please mod parent down by Anonymous Coward · 2003-08-23 02:40 · Score: 0

Where do you get the opinion that a 5.5 giga flop machine costs 199 $?

The Apple machine is a super computer but at a cost of 1940 $/Gigaflop. Single machines cost a lot.

How you get the value of 199 $ is beyond me.

You are very uniformed and shooting from the hip.

Re:Please mod parent down by Anonymous Coward · 2003-08-23 02:45 · Score: 0

Try reading the post again.
Re:Please mod parent down by gorim · 2003-08-23 02:52 · Score: 2, Informative

A playstation2 costs $199. That information is in your local newspaper. Actually, sales peg it at $179 lately, my mistake. The playstation2, with 2 vector processing units, each with 4 floats wide registers (128bit), capable of doing a multiply-add operation per clock cycle on whole registers, at 300mhz independant of the main CPU which still has its own scalar floating point coproc, handily does 5.5GFLOPS, and is well documented as such if you google around. Check out http://playstation2-linux.com/
Re:Please mod parent down by Bytal · 2003-08-23 02:56 · Score: 1

The article does address this if you read it.

"...An even cuter comparison is with this, a $50,000+ system built using 70 PlayStation2 units. Not only does KASY0 have a vastly superior network and significantly higher floating point performance per node (8 GFLOPS vs. 6.5 GFLOPS for the PS2), but we get LOTS more nodes!..."
Re:Please mod parent down by crazysim · 2003-08-23 04:13 · Score: 0

You forgot you gotta buy the kit!!!

So much power... by krahd · 2003-08-23 02:41 · Score: 5, Funny

and it still can't run Doom III at a decent rate.

--krahd

mod me up, scottie!

--
mod me up scottie!

I wonder... by MoronGames · 2003-08-23 02:41 · Score: -1, Redundant

I wonder how many frames per second it gets in Quake 3!

--
hey!

The Mare by Anonymous Coward · 2003-08-23 02:42 · Score: -1, Offtopic

A full grown stallion's cock, when fully erect, will measure some two to
three feet long. It can be three to six inches thick at the base, to about
two inches thick at the head. Horses are somewhat different from other
animals in the way their cock head works. When a horse is fully erect and
excited and ready to mount, his cock head is somewhat pointed and not as
thick as might be normally observed. This is to facillatate an easier
entry into the mare. After the horse has entered and reaches a climax the
head swells (though it is more spongy then hard) into a fist sized mass as
he ejacultates. It is thought that this serves as a plug to force the
semen deep into the mare rather then allowing it to leak out. A full grown
stallion can ejaculate about one cup ( 8 ounces ) of semen. It will take
quite a few spurts to accomplish this. Each time his tail will raise and
lower in a brief flick. The first few jets are of a thin to average
consistency of cum. The final few jets are of a thick gelatinous
substance... it is thought that this serves to "seal" the mares pussy so
that the semen has time to do it's thing before leaking out. Horse semen
is extremely viscous, if you touch your finger to a pool of it you can draw
a thin string of it five to six feet long! Horse cum has a nice flat taste
to it...not at all bitter like man's cum. You can easily drink cups of it
with no discomfort.

The Mare - how to do it.

Mares can be quite satisfactory for the average well endowed male. If you
are somewhat less developed you might find better pleasure with a pony or
Miniature Horse. These are also better as they are lower to the ground. A
pony you can fuck standing up. A miniature horse on your knees or
squatting depending on the size. A mare will require something to stand on
or "platform shoes"...(IE mini stilts to raise you a foot off the ground)
so that you can reach her pussy.
Fucking any horse will depend on the horse. Some will be ready right
away...some will take coaxing. Pet the animal, talk to it softly, spend
time with it gaining it's trust. If something you are doing upsets it then
don't force it. Talk to it and calm it. If you work slowly you can make
an animal accept anything. It is just a question of helping it overcome
it's fears. All animals fear man if raised in the wild. How any animal
reacts will depend on it's own experiences. If you haved raised the animal
yourself in a loving enviroment, then you should have no problem
associating with it, if it is a strange animal that you have met in the
wild then you will have to go through an extended "courtship" to learn how
to respond to the beast.

MARES - TRAINING YOUR OWN

When the filly reaches weaning age, seperate her from her dam. If you have
limited time to spend then she should be put to pasture. If you have
plenty of time then you should keep her in a stall. Spend time with her
during the day petting and grooming her and allow her some time to run
free. Limit her access to other horses though and see that she spends at
least 8-12 hours a day in the stall. (Start with more free time and as she
approaches her first birthday confine her more...she is now at the right
age and her confinement will have made her so bored that she is amenable to
any new experience so long as it is not unpleasant)Young fillys have no
objection to someone playing with their pussy's. I have walked up on a pen
full of strange fillys at night and they came right up to me and I petted
them and felt up their pussys and they just lifted their tales and seemed
to enjoy it. These fillys didn't even know me but they were young,
inexperienced and bored...also since they were penned they were used to the
presence of people and did not fear me. Most horses in a large pasture
will run when they scent a strange human in their pasture at night.
If you sit on the ground and wait patiently, they will get downwind of you
and s

/. monitoring your activity! by Anonymous Coward · 2003-08-23 02:43 · Score: -1, Offtopic

Look at the page source and the small 1 pixel square under the ad!

iframe width=1 height=1 border=0 frameborder=0/iframe

This is creepy! What are they doing with this information? Selling it to Microsoft?

cable management by HBI · 2003-08-23 02:44 · Score: 3, Interesting

What a mess of cables! I understand they were hitting a price point, but would it have killed them to spring $500 or so for a cable management system?

There's something professional looking about having the cables look neat. On the other hand, maybe i'm just anal about things.

--
HBI's Law: Frequency of calling others Nazis is directly correlated with the likelihood of the accuser being Communist.

Re:cable management by Anonymous Coward · 2003-08-23 03:26 · Score: 0

You don't even need to spend a lot of money. A few bags of velcro cable ties, a lable printer and a cat5 tester will save you an incredible amount of hassle in the long run.
Re:cable management by imsabbel · 2003-08-23 04:09 · Score: 1

The main point of the network is a semi-random optimized nodo2node network using 2 parallel nets.
It would be almost impossible to use a cable management system without creating a greater mess.

--
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
Re:cable management by HBI · 2003-08-23 04:16 · Score: 2, Informative

I would consider something like this, bent into a circular pattern.

We use this in one of my closets to provide a feed from the ceiling to a rack with a patch panel. We had moved the rack 4' to the left and we needed something to bring the ceiling stuff down to the new rack. We used fittings with tek screws and wire ties to hold it firmly in place and linked it into the drop ceiling framework as well for additional stability.

I could see doing something like that here between the stainless shelving units. Other adequate solutions could be arrived at. Also, the snake tray probably cost us $150 tops.

--
HBI's Law: Frequency of calling others Nazis is directly correlated with the likelihood of the accuser being Communist.
Re:cable management by ModernGeek · 2003-08-23 07:15 · Score: 1

They said they could have made the cables alot neater, but they were wanting to make it so that the cables can be changed over time since it's for research, it will be moved around and updated alot, so they wanted to keep it loose.

--
Sig: I stole this sig.
Re:cable management by evilviper · 2003-08-23 17:08 · Score: 1

There's something professional looking about having the cables look neat. On the other hand, maybe i'm just anal about things.

From personal experience, I can attest that oragized cables save a great deal of time. I would estimate that it takes about 10x longer to replace an ethernet cable in a mess of cables, than it does when they are neatly organized (and labeled or color-coded).

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant

Now just imagine... by Anonymous Coward · 2003-08-23 02:45 · Score: -1, Redundant

The Beowulf cluster of money needed to pay the Beowulf cluster of SCO licenses to avoid the Beowulf cluster of SCO lawyers pursuing these guys to bite their asses.

100 SERIES by Magic+Thread · 2003-08-23 02:45 · Score: 1

Cooling by bengoerz · 2003-08-23 02:45 · Score: 4, Informative

I toured the previous cluster these guys did (KLAT2) and was very impressed. However, using AMD Athlon Thunderbirds last time, it did get quite hot. I remember standing by the cluster looking at all the wiring and being bombarded by an overhead cooling vent. I'm also assuming that these cooling issues is the reason that each case has two blow-holes. I'd also like to see these guys post in-depth specs of each machine. Being a hardware nut, I'd like to see how they got so many machines so cheap, and maybe even what vender they used. As I remember, they worked REALLY hard on their last cluster to keep costs to an absolute minimum.

Re:Cooling by imsabbel · 2003-08-23 03:07 · Score: 1

Well, enough cpus in a room, no matter what type, will generate a lot of heat.
And price:
http://aggregate.org/KASY0/cost.html
It should be easy the extrapolate the system configs :)

--
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
Re:Cooling by evilviper · 2003-08-23 17:01 · Score: 1

Cooling is a point I was going to make myself... Although quite a bit less kind than your post.

Looking at the photos, it looks like they just stuffed a dozen towers on a rack, and plugged them in. Needless to say, this is a TERRIBLE way to do things, and no doubt they are going to be paying many times the cost of the systems in the price of cooling that room.

What bothers me is how often I see things like this happen, and how dirt cheap a little bit of ducting is... If they would have put just a few more dollars into it, they could have ducted 99% of the heat out of the building, saving a great deal of money on cooling in the long run, and probably keeping the system alive longer, since they won't be sucking in their own heated exhaust.

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Re:Cooling by bengoerz · 2003-08-24 02:39 · Score: 1

In fairness to them, their Air Conditioning is so strong that very little of their exhausted heat is making it back in to the systems. They've probably got several thousand CFM of cold air pumping down on these things. Another thing to keep in mind with their setup is that they are in a University building, and I don't think they have to pay the utilities. Why worry about power issues when the University will take care of it for "free"?

Hmmm. by Pig+Hogger · 2003-08-23 02:45 · Score: -1, Redundant

If supercomputers are already beowulf clusters of sort, I wonder what a beowulf cluster of those would be...

Is that a real number or a marketing number? by Sycraft-fu · 2003-08-23 02:48 · Score: 5, Insightful

I'm guessing the latter. You see all sorts of BSified numbers from marketing departments on processors, but they have little to do with reality. The number for this AMD cluster is a real, actual, measured-using-a-real-world-app number. To give you some idea of BS console numbers, the Xbox has a PIII 733 processor in it (ok, technically it's a little different, but it's a P3 core). Now the Gflop claim is 2.93. Out of a P3 733? Ya right, on paper perhaps but never in the real world, much less on a real app.

Then, of course, there is the issue of specialised chips vs normal chips. A GeForce 4 4400 can claim, roughly, 80 Gflops peak. That sure beats the hell out of any sinlge CPU I've ever heard of, including the Power4. Thing is the GeForce 4 is a graphics DSP, it isn't a general purpose CPU. It can do that kind of math when all its units are working at what they do best, but try to reprogram it to do something else and it will slow to a crawl (for that matter I'm not even sure that it is turing complete).

So don't take any hype on a console to equate to real performance in a general task. Oh, and the BS marketing number I see for the PS2's Emotion Engine is 6.2Gflops.

Re:Is that a real number or a marketing number? by drinkypoo · 2003-08-23 04:41 · Score: 2, Insightful

Of course, assuming it's only half the parent comment's assertion, thus 2.25 GFlops, at $180 it's still cheaper than $100/GFlop. However, as others (should?) have pointed out by now, it's useless as a supercomputing node for all but the smallest tasks since it has no local storage and extremely limited main memory. You will have to spend another $200 for a linux kit to get storage and networking, bringing it up to $380 for the system. If it were actually 5.5 GFlops in the real world, then that would still be cheaper than $1/GFlop, but of course neither you nor I believe it is that fast while doing general-purpose processing.

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:Is that a real number or a marketing number? by Sycraft-fu · 2003-08-23 12:28 · Score: 1

I bet it is less than half. Those BSified numbers tend to be the ultimate best case, on paper situation. Generally speaking, they are how fast the chip can theoritically perform its fastest operation(s). Well right off the bat you loose speed going from theoritical to real, nothing ever performs 100% of theoritical. Then there is the problem of memory access. All these theoritical numbers are assuming just ops. Great, doesn't work like that, you have to hit memory to get data. MAJOR slowdown there. Then there is the fact that not all your ops perform as fast as your fastest one. This goes double for a chip like the EE that has lots of vector shit related to graphics. So you aren't going to see that kind of speed general purpose. Finally there is the issue of supporting instruction, like load/store fom memory and so on. Need to take those into account in additon to the time to access memory itself.

In reality, I'd be supprised if the EE can do 1 Gflop on general purpose tests. I'm sure it does more than that playing games, when it is being used to do graphics (the PS2 is weird, the EE is half CPU, half GPU), but not for general purpose calculations like Ray Tracing.
Re:Is that a real number or a marketing number? by Jeff+DeMaagd · 2003-08-23 15:40 · Score: 1

http://aggregate.org/KASY0/faq.html

What performance does KASY0 really get?

It really gets over 471 GFLOPS on a 32-bit version of HPL. Using an "untuned" 64/80-bit version, KASY0 gets a very respectable 187.3 GFLOPS. These aren't theoretical numbers, they are the real thing. The theoretical we-will-never-see-that numbers are 531 GFLOPS and 1.06 TFLOPS, respectively, for 64/80-bit and 32-bit floating point.

Yes, we know HPL is only one application and not a very general one at that. We have other stuff running as well... but most of what we do is computer system design. Thus, our primary applications tend to be things like the Sparse Flat Neighborhood Network design program, which nobody else yet has. The result is that the performance numbers that are most important to us are meaningless to anybody else.

There is a lot of info in that FAQ.

Re:Playstation2 at 5.5GFLOPS costs only $199 $40/G by Surak · 2003-08-23 02:49 · Score: 1

5.5 Gflops, I dunno if it can really do that, but ...uh..the point is that it's the first *supercomputer* to break the $100/GFLOP barrier. The Playstation2, last I checked, isn't a supercomputer, it's a videogame platform.

--
My journal has hot /. gossip.

Here is the bill! by borgdows · 2003-08-23 02:50 · Score: 5, Funny

Dear customer,

At the cheap introductory price of 699$ for 80 lines of code in the Linux kernel, it will cost you 8,377,500$ by kernel since we have discovered that in fact 1000000 lines of SCO IP were copied into Linux.

Designation .. Price .. Qty .. Total
Linux kernel .. 8,377,500$ .. 128 .. 1,118,400,000$

So you must pay us only 1,118,400,000$, and in my kind almighty I will offer you a discount of 118,400,000$ so you only have to pay ONE BILLION DOLLAR if you pay before tomorrow!

Please send you creditcard number at darl@sco.com

Sincerely yours,

-- Darl Mac Bride

Re:Here is the bill! by McBride,+Darl · 2003-08-23 02:55 · Score: -1

Please note that the parent is not in fact Darl McBride! This poser has stolen my online identity and I will be contacting my lawyers about this matter shortly.

--
Darl McBride
Chief Executive Officer
Caldera International, Inc.
Re:Here is the bill! by Anonymous Coward · 2003-08-23 02:58 · Score: 0

FUCK YOU DARL!!
Re:Here is the bill! by drooling-dog · 2003-08-23 05:35 · Score: 1

You have lawyers?

How long until... by r00zky · 2003-08-23 02:50 · Score: 1, Funny

- Casio sues them for trademark infringement?
- SCO asks them 92268$ worth of licenses and/or sues them for copyright infringement?
- KFK sues em for patent infringement? (you know, "the method to fry one billion chicken flaps (GFLAPS) for under 100$ with AMD processors")

Ah the USofA, Land Of the Lawyers.

--
I'm a chainsmokin' alcoholic sociopath, so-ci-o-path

Where are the hard drives? by BubbleNOP · 2003-08-23 02:51 · Score: 1

I didn't see hard drives on their parts list. Why is that? How do they boot them up?!

Re:Where are the hard drives? by borgdows · 2003-08-23 02:53 · Score: 1

They boot up on a USB key! ;)
Re:Where are the hard drives? by zeroclip · 2003-08-23 03:05 · Score: 1

This is where they cheated :P It says on the page that they boot off a another cluster.
Re:Where are the hard drives? by martman00 · 2003-08-23 04:00 · Score: 0

Netboot. They have a tftp server and they have ethernet cards that support netbooting. All you need to do is pop in another chip.

Nice wiring! by nate.sammons · 2003-08-23 02:53 · Score: 2, Insightful

Looks like most of the wiring jobs I've seen done by students: kasy0core.jpg.

God forbid they use cable gutters ;-)

Other than that, kick ass job guys!

-nate

Re:Nice wiring! by KingDaveRa · 2003-08-23 03:15 · Score: 1

And what ever happened to colour-coordination? Nothing matches!

In cache maybe by msgmonkey · 2003-08-23 02:55 · Score: 2, Informative

These numbers for microprocessors etc mean nothing because they are usually referring to operations on data in cache.. you'ill find that real life performance is 10-20x slower because thats how much slower accessing main memory is.

From the KAYSO document... by FreeLinux · 2003-08-23 02:55 · Score: 1

Note that every PC has at least one single-switch-latency path to every other PC; some PC pairs have more than one such path.

Every host does have at least one pathe to every other host but, most hosts have multiple paths to other hosts. It is true however that all hosts do not necessarily have multiple paths to all other hosts.

Way to go! by panda · 2003-08-23 03:00 · Score: 1

Hey! I used to work there.

Way to go Dr. Dietz!

So, mod me anyway you want, karma to burn.

--
Just be sure to wear the gold uniform when you beam down -- you know what happens when you wear the red one.

Re:Playstation2 at 5.5GFLOPS costs only $199 $40/G by alienw · 2003-08-23 03:00 · Score: 1

Nice how you take the numbers from a marketing press release and treat them as if they are the absolute, indisputable truth. Can you show me the actual, reproducible benchmark that produced those numbers?

Also, the PS2 is not a supercomputer. It has a slow processor and very little RAM, so it wouldn't be able to do much number-crunching. You can't hook PS2s together, anyway, so comparing a single specialized machine to a cluster is absolutely meaningless.

Hmm Math? by Anonymous Coward · 2003-08-23 03:01 · Score: 0

128+4...

That's like 132 isn't it?

Re:Hmm Math? by borgdows · 2003-08-23 03:02 · Score: 0, Redundant

woww!!
Do you have used a supercomputer to calculate this impressive formula??
Re:Hmm Math? by r00zky · 2003-08-23 03:17 · Score: 2, Informative

128+4...
That's like 132 isn't it?

From the FAQ:

KASY0's configuration is:
128 + 4 "cold spare" PC nodes, each containing:
One AMD Athlon XP 2600+ (the 2.075GHz version)
One 512MB PC2700 DDR SDRAM
BioStar M7VIT Pro motherboard
Two Linksys LNE100TX NICs
Codegen 6042L case with 400W power supply
18 BenQ SE0024 24-port Fast Ethernet switches
405 Cat5 Fast Ethernet cables
RedHat Linux 9.0, modified Warewulf 1.11

So it's 128, the other 4 are spares!

--
I'm a chainsmokin' alcoholic sociopath, so-ci-o-path
Re:Hmm Math? by DirkDaring · 2003-08-24 02:37 · Score: 1

The best supercomputer in history, his brain. Or at least I would hope thats what he used.

Wrong by imsabbel · 2003-08-23 03:02 · Score: 3, Informative

In reality, beowolf clusters are good for only a subset of supercomputing tasks and the "real" supercomputers are still best at general purpose supercomputing.

If you can paralize your application well enough, beowoulf rules, but if you need a lot of node2node communication, the network cost quickly surpasses the cpu cost of the system

--
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?

Re:Wrong by glueball · 2003-08-23 03:24 · Score: 1

Mod this up.

imsabbel is absolutely correct.

Bill
Re:Wrong by sjames · 2003-08-23 06:23 · Score: 3, Insightful

Really, it's a spectrum. One one end you have fully commodity beowulf, in the middle, you see things like Dolphin and Myrinet, and on the high end you see fully custom backplanes and sometimes RAM and I/O controllers as well. Purpose built CPUs are becomming less common now, but not unheard of.

Each step up the spectrum widens the domain of problems that the machine can work on efficiently, and raises the price for the machine. In many cases, a 'real' supercomputer is more or less a cluster with a specialized network and OS and mounted in a single cabinet so it doesn't look like a cluster.

In general when a lower end machine can efficiently run your program, there is no benefit to using a more expensive machine.

As server hardware improves and 'exotic' hardware becomes more mainstream, the gap between the low and the high end narrows. There will probably always be a small but existant set of problems that call for the 'real' supercomputer, but that set is shrinking.

There are other considerations as well. If the Beowulf in your lab can solve the problem in 1 week and is available now, while the 'real' supercomputer on the other campus can solve it in 4 hours and will have a timeslot available in 2 weeks, the Beowulf is 'faster' from your point of view.

yeah, you like to fuck men by Anonymous Coward · 2003-08-23 03:02 · Score: -1, Troll

don't you? you gay little weiner.

How many university have larger clusters? by SilverSun · 2003-08-23 03:04 · Score: 5, Interesting

I wonder which universities/institutes have larger and maybe cheaper clusters, but just don't bother with running benchmarks. I for one are sitting next next to a tiny cluster with 40 dual-cpu nodes, which is connected (GRID like) to a 340 dual-node cluster in a nearby town. Non of us high ernergy physicists bothers with running any benchmarks on our clusters, other than our own applications. I wonder how many "linux-cluster-supercomputers" are out there which would easyly make it into the top 500, but noone has ever heard of....

Cheers.

--

KdenLive/PIAVE - non-linear video editing

Re:How many university have larger clusters? by Anonymous Coward · 2003-08-23 03:16 · Score: 1

Right on. I am a sysadmin in a uni Physics dept. and we have some serious parallel computing going on (our beowulf has 95 dual p4 compute nodes on 4 gbit switches) but never bother with benchmarking. Plus, out of hours just about every linux box in the place becomes a node of a big open mosix cluster. Maybe we're just too busy to stroke our own egos...
Re:How many university have larger clusters? by Chilles · 2003-08-23 03:28 · Score: 3, Funny

I wonder how many "linux-cluster-supercomputers" are out there which would easyly make it into the top 500, but noone has ever heard of....

Well... probably more than one, definitely no more than 500.
Re:How many university have larger clusters? by imsabbel · 2003-08-23 04:06 · Score: 1

At my university, there isnt a "real" cluster or supercomputer ;(.
But there is a gigabit connection to the 1.6 TFlop Hitachi in Munchen...

--
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
Re:How many university have larger clusters? by Anonymous Coward · 2003-08-23 08:01 · Score: 0

Same here. I work for a private biotech sysadmining 2 clusters( 500 nodes dual p3's running openmosix) and not benchmarked. To me as long as my customers can get the work done, I couldn't care less what it's linpack score is. I judge it's sucess more by the results achieved and what good it is doing for the company and science in general.

When I have scientists thrilled that they have discovered something that wasn't possible without the resources I support that is more than enough reward.

I also know of at least 2 other private companies here in Socal that have at least the same or more resources without a press release or submission to the TOP500.

Dance MP3 Needed! by Anonymous Coward · 2003-08-23 03:06 · Score: -1, Offtopic

Where can I find the Who Da Funk's "Shiny Disco Balls"? Or Grand Popo Football Club's "Men Are Not Nice Guys"?

TIA

University students by SuperBanana · 2003-08-23 03:07 · Score: 5, Insightful

Because this was a university project, KASY0 was assembled entirely by unversity students, which while being a source of cheap labor, is also a good way to get a lot of students of involved in a great project.

At the risk of being flamebait- No. Using university students is almost always purely a way of getting cheap labor to do semi-mindless, or completely mindless, stuff the staff doesn't want to do- it's a common myth that students 'learn' by doing grunt work. I should know- I have several grad student friends, and they've thusfar spent a large part of their academic careers working in labs doing mind-numbingly boring stuff(according to them.)

Imagine if a Bio lab did this. The following would sound pretty absurd: "Help us move our lab, you'll learn about cellular recombination!". No. You'll learn what a bunch of lab equipment looks like, how eccentric the professors are, and how expensive/fragile/heavy the equipment is, and the next morning what sore muscles are like. Let's get a reality check here.

(from the site):Our group develops the systems technology for cluster supercomputing; the more people we can show how to apply these technologies, the better.

Huh? What cluster supercomputing "technology" does assembling a PC and plugging it into ethernet teach you? Did they give a presentation about how clustering technology works, for example? Did they explain to each person, as they put a machine in a particular place and wired it to a particular switch, WHY it was going there etc? Obviously I wasn't there, so perhaps someone from the group can contribute on this point.

--
Please help metamoderate.

Re:University students by panda · 2003-08-23 03:46 · Score: 4, Informative

Having worked there, and knowing what Hank Dietz and his students are doing, I can tell you that it is different from just slapping PCs together, stringing wire between them and installing clustering software.

Dietz specializes in networking and all the wiring that you see in the photos is charted out by custom software that he's written just for this purpose.

He works in the realm of optimizing communications among the nodes to avoid network latency and so on. If you read the POVRay benchmarks, you'll notice that the author comments that several clusters' CPUs spend most of their time idle due to network latency. Dietz is researching the best ways to eliminate much of that latency so that the CPUs in the cluster can spend more of their time crunching data rather than just throwing off heat. To my knowledge, he is succeeding at this and better than most other researchers in the field.

As for what his students learned from this, I don't know exactly which students helped him on this. For KLAT2, there were several undergrad volunteers who helped with wiring and assembly, mostly from the campus Linux Users' Group. I know his grad students and research assistants are learning a lot about how clustering and network tech works, and a couple are doing their Ph.D. disserts in this very subfield of E.E.

--
Just be sure to wear the gold uniform when you beam down -- you know what happens when you wear the red one.
Re:University students by drinkypoo · 2003-08-23 04:44 · Score: 1

It seems to me that the next step is to get some big switches with VLAN support and reconfigure them dynamically as the workload changes in order to maximally utilize all nodes. I wrote some pathetic little software once to log into some switches and make vlan changes from a web interface (no security or anything, what a bad idea eh? worked though. this was before cisco included the "why bother" ssh1 in ios) so at least THAT part is trivial :D

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:University students by Anonymous Coward · 2003-08-23 05:08 · Score: 0

mod parent down that's a long reply that's off-topic he dodges the issue raised by its parent -- that the volunteers aren't learning anything, which they're not...he wasn't talking about the grad students who work for Dietz basically you get a bunch of undergrads fired up about something because they're uninvolved/don't know any better and throw in the few optional pizzas...that's how it usually happens, and I happen to know for a fact that's how it happened here it's all about cheap labor
Re:University students by Kethinov · 2003-08-23 05:11 · Score: 1

At the risk of being flamebait- No.
He who moderates you has been infected with the reverse psychology bug! See sig.

--
You're right, I wouldn't steal a car. But if it were possible, I sure as hell would download one!
Re:University students by rapett0 · 2003-08-23 06:34 · Score: 1

I worked there as well back during the KLAT-2 days. Sure I dont' remember getting any actual mention on the project (even if I did help coordinate the student help, help build it, assemble it, etc), but thats ok. It was a great project to work on (I still even have the GaLUGtica videos I made in 3dsmax) and I did learn quite a few things. Deitz harpy of darkness, aka Tim Mattox, knows his stuff, and him along with Dieter will very helpful to the students answering whatever questions they had.

Petty 1 of 2.
Re:University students by Skuld-Chan · 2003-08-23 07:29 · Score: 1

On the other hand to build a supercomputer for less than 89$ per gflop you still have to actually "build it". I mean who else is going to put it all together? Someone has to build it if you want cs students to use it.

And yes microbiology students will still have to build their own apparatus for experiments they conduct - I only know this because I took a class in microbiology a while back and I had to build the apparatus for all the required experiments I had to do.

I'm guessing in this case they not only put it together, but at the same time are going to use it and be trained in how super computers like this work.
Re:University students by evilviper · 2003-08-23 16:57 · Score: 1

Imagine if a Bio lab did this. The following would sound pretty absurd: "Help us move our lab, you'll learn about cellular recombination!". No. You'll learn what a bunch of lab equipment looks like, how eccentric the professors are, and how expensive/fragile/heavy the equipment is, and the next morning what sore muscles are like. Let's get a reality check here.

Well, if they had their computer science students doing it, it wouldn't be of any benefit to them, but if it was being done by some sort of PC Repair class, it certainly would be benefitial to the students.

Also, who's to say that they didn't go into further detail. One of things you do in just about all embedded systems courses is to build the circuit yourself, then make it work. The building of the circuit would be considered grunt work unless you later program it to do something yourself.

Yes, a lot of schools have a lot of students doing crap work, but I'm not going to condem a school for a situation like this, where I don't know enough about it to make such a judgement.

Do you know something the rest of us don't? Do you know that the process was not benefitial to the students, or are you just ranting?

(This is from someone who worked at a college for several years, saw first hand both extremes, and had many students doing many similar things)

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant

And how much HEAT? by Surak · 2003-08-23 03:08 · Score: 2, Funny

I mean these things are Athlons! Heck, they're saving money just from the fact that they'll never have to turn on the furnace again!

Did you guys notice from the pics that there doesn't seem to be any fans in the holes on the sides? Are they crazy? These are Athlons. I hope they put enough fans in those things.

--
My journal has hot /. gossip.

Re:And how much HEAT? by lederhosen · 2003-08-23 03:18 · Score: 2

As much heat as power.
Re:And how much HEAT? by Anonymous Coward · 2003-08-23 04:03 · Score: 0

Give me a break, Intel is trying to make a new case standard where there is a case fan blowing out the side directly above the CPU.

P4's get hot too, and the whole "every Athlon is super hot" is an urban myth.
Re:And how much HEAT? by ptbarnett · 2003-08-23 04:47 · Score: 1

Did you guys notice from the pics [aggregate.org] that there doesn't seem to be any fans in the holes on the sides?
See here:
For example, each case came with two side fans, which we converted into a redundant stack venting out the back.
Re:And how much HEAT? by drinkypoo · 2003-08-23 04:51 · Score: 1

My Athlon XP 1700+ overclocked to 2000+ dissipates approximately 46W as heat. With cooling moving only from the front of the case to the back (though including one pretty fast and loud fan) it reliably stays below 104 degrees Fahrenheit. I'm guessing they're not using overclocking, and they're using the new CPUs with the higher speed bus. ZDNet claims that the 2600+ dissipates 62.0 watts as heat, so there's a bit of a bump there, but since I know from experience that Athlon chips can run at 140 degrees without trouble (my prior athlon-tbird overclocked from 1.33 to 1.4 and in a hot house) I don't think they have much to worry about.
By the way, the holes in the sides won't help there, sparky. Only the case on the end would derive any benefit. I'd actually worry, though, about the case on the end overheating due to improper airflow since it's just open. I'd be slipping a manila folder in there and taping it into the inside of the case, unless I got uppity enough to use a metal plate and rivet it in. (unlikely.) Most of the time PCs have room for a fan in the bottom front area, where full-length cards would reach if you had one (almost no one does any more.) Then the air will flow upwards through the case and out through the power supply, if the fan blows the proper direction anyway. (If it doesn't when you get it, well, I've turned quite a few of them around.) It makes NO sense to have your power supply heat up the air and blow it into the computer, just like it makes no sense for the water in a car to flow through the heads last, even though that is precisely the way most engines do it. (It was a big deal when the corvettes got an engine with "reverse flow cooling" for example.)

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:And how much HEAT? by afidel · 2003-08-23 05:19 · Score: 1

The Athlon actually has a pretty average Watt/Flop ratio for a modern processor. The only one that really trounces it is the POWER series including POWER-PC. The Athlon 2600+ only uses 68.3W, compare this to a 2.4Ghz P4 which uses 66.2W and you see that they are in the exact same neighborhood. And if you include price into the equation the Athlon becomes the leader. Also if you had RTFA they explain that the side fans were moved to a stacked rear configuration for better airflow and redundency.

--
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
Re:And how much HEAT? by evilviper · 2003-08-23 17:17 · Score: 1

I mean these things are Athlons! Heck, they're saving money just from the fact that they'll never have to turn on the furnace again!

I'm getting exhausted just posting this over and over on /., but here we go, one more time...

The heat output of P4s is significantly MORE than that of an (equivalent-performance) Athlon XP. In addition to the P4 putting out more heat, the P4 has a LOWER maximum operating temperature.

--
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant

In other news... by rmdyer · 2003-08-23 03:09 · Score: 1, Insightful

Now that the university students have graduated and moved on, there isn't any documentation, nor do they know how to use the darn thing...

-1

why not DSP? by mike_g · 2003-08-23 03:11 · Score: 4, Interesting

Why are not DSPs used in configurations such as this. The TI 67xx series are able to perform about 1 GFLOP/s running at only 150 MHz and cost only about $40 per chip.

This price/performance ratio seems to make them very attractive compared to general purpose CPUs. According to the NASA G5 Study, the P4 2.66 GHz is only able to achieve 255 MFLOP/s. And the P4 costs about 4x the price of the 6711 DSP.

It seems that DSPs should be the clear winner in supercomputer applications, what are their disadvantages and why are they not used? Granted there is a lack of mass produced hardware such as motherboards for DSPs, but that alone should not exclude them from the supercomputer realm.

Re:why not DSP? by latroM · 2003-08-23 03:35 · Score: 1

Maybe because DSP is designed for digital signal processing, not to handle branches and stuff that ordinary CPU does.
Re:why not DSP? by panurge · 2003-08-23 03:40 · Score: 1

I guess it's the lack of operating system support on the DSPs themselves. Plus their instructions sets and I/O don't lend themselves well to general purpose computing. The cost of developing a node consisting of a DSP plus a general purpose processor, plus the efficient I/O to the DSP, might be too high for the relatively restricted usage on supercomputers.
That said, my Palm Tungsten is a combo of a GP processor and a DSP, as I believe are several Sony variants. Perhaps as I/O on handhelds improves (?) the time will come when a university produces a Beowulf cluster of handhelds and the world's smallest supercomputer per gigaflop (more like ly teraflop by then).
In case this is more than just my usual totally uninformed wild speculation, could I register claim to the idea. In order to obtain a (free) license, just name the first model Grendel's Bane. And promise not to make any SCO jokes.

--
Panurge has posted for the last time. Thanks for the positive moderations.
Re:why not DSP? by mike_g · 2003-08-23 03:50 · Score: 1

Maybe because DSP is designed for digital signal processing, not to handle branches and stuff that ordinary CPU does.
True, I realize this. But I am under the impression that a lot of heavy duty number crunching algorithms have a minimum number of branches, and mostly just perform the same operations on multiple sets of data. Think of the FFT and simulations of systems based on differential equation models. This should include weather models and quite possible nuclear events. These applications seem like areas where the DSPs would shine and where supercomputers are most often applied.
Re:why not DSP? by girouette · 2003-08-23 04:23 · Score: 1

For weather modelling, I can perhaps envision DSPs as co-processors, but much of the modelling work would still have to be performed by general-purpose CPUs.

Pure atmospheric dynamics can be done in the spectral domain, but other physics (convection, cloud physics, radiative processes, surface/atmosphere interactions) are handled by algorithms that usually require more than just solving a transform or a differential equation.
Re:why not DSP? by SmackCrackandPot · 2003-08-23 04:36 · Score: 3, Informative

Actually, they do, but they are referred to as vector processors rather than DSP's. Probably the most famous and the first was the Cray supercomputer. And there was also the INMOS "Transputer"

DSP's are optimised to handle streamed data of a particular maximum size (Eg. 4-element float point variables). Useful for image processing (red,green,blue,alpha) and 3D graphics(XYZW), but if you're modelling something like ocean currents, global weather, every data element is more than likely going to have more than four variables (eg. temperature, humidity, velocity, pressure, salinity, ground temperature), you may not get full optimisation.

Plus, you also need a means of getting all these processors to talk to each other. DSP's are nearly always optimised to operate in single pipelines, so don't need much communication support (eg. Sony Playstation 2). However, if you're designing a supercomputer system, the major bottleneck is the communication between processors (network topology). Some applications might only need adjacent processors to talk to each other (global weather simulation usually represents the atmosphere as a single large block of air, with sub-blocks assigned to seperate processors. Other applications might assign individual processors to different tasks, which complete at different rates (eg. the Mandelbrot set). A configurable network architecture allows the system to be used for many more different applications.
Re:why not DSP? by sigwinch · 2003-08-23 09:07 · Score: 1

Why are not DSPs used in configurations such as this.
1. Non-commodity hardware has high one-time expenses for design.
2. DSPs tend to not have a lot of RAM, whilst big modelling apps crave RAM (esp. raytracing).

--
--
Kuro5hin.org: where the good times never end. ;-)
Re:why not DSP? by AtrN · 2003-08-23 09:55 · Score: 1

The transputer was not a vector machine, not unless you count the move2d instruction :)
Re:why not DSP? by SmackCrackandPot · 2003-08-23 12:30 · Score: 1

My mistake. I was thinking of the dual combo system of the TTM220 which combined a Intel i860 and a T805.

Congradulations to them by Elpacoloco · 2003-08-23 03:24 · Score: 1

Yeah, good work.

What will they do with all this processing power? Farm it out? Boast? Serve a pr0n company?

Can't hook ps2's together? by way2trivial · 2003-08-23 03:24 · Score: 1

then what is this little box for?
http://www.us.playstation.com/hardware/networkadap tor/SCPH-10281.asp

--
every day http://en.wikipedia.org/wiki/Special:Random

Re:Can't hook ps2's together? by alienw · 2003-08-23 04:45 · Score: 2, Interesting

Good luck getting a beowulf cluster with that crap. Ethernet is not a good interconnect technology. It's not even a good networking technology. And interconnect technology is the main performance-determining factor with a beowulf cluster.

Anyway, if you think you can do better with PS2s, why don't you do so?
Re:Can't hook ps2's together? by way2trivial · 2003-08-23 22:43 · Score: 1

I don't think I can,
but an argument/post should contain valid info- or it's less than with consideration.. I knew ps2's could be hooked together, and I don't own a console.. not everyone does know that.

--
every day http://en.wikipedia.org/wiki/Special:Random

Mods on crack by imsabbel · 2003-08-23 03:25 · Score: 1

"Insightful"?!?
Yeah, right.
Guess everybody from universitys to big corporations spend tons of money for supercomputers/clusters because they have no use for them.
Sounds very likely.
Given the fact that your other posts in this topic all concern the blackout or McDonalds->
MOD PARENT DOWN (OFFTOPIC)

--
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?

Mckenzie Cluster, faster, cheaper per TFlop by prof_bart · 2003-08-23 03:33 · Score: 5, Interesting

Hmmm...

Nice machine, but this January, CITA and the astro department at the University of Toronto brought a 256 node dual Xenon system on line: "1.2 trillion floating point mathematical operations per second (Tflops) on the standard LINPACK linear algebra benchmark." Total cost: CDN$900K (including tax) (in January prices, that's $600K U.S. or $0.50USD/GFlop.) It's being used for some very cool Astro simulations...

See http://www.cita.utoronto.ca/webpages/mckenzie

Re:Mckenzie Cluster, faster, cheaper per TFlop by Anonymous Coward · 2003-08-23 04:28 · Score: 0

Actually, that's $500/GFlop.
Re:Mckenzie Cluster, faster, cheaper per TFlop by sracer9 · 2003-08-23 05:24 · Score: 1

That's pretty decent. And I bet it was a gas to setup.

--

No thanks. I don't smoke anymore.
Re:Mckenzie Cluster, faster, cheaper per TFlop by afidel · 2003-08-23 05:41 · Score: 2, Informative

$600,000 dollars / 1,200 Gflops = $500/Gflop. I think you misplaced a decimal in the $Cad->$US conversion =)

--
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
Re:Mckenzie Cluster, faster, cheaper per TFlop by Dahan · 2003-08-23 11:19 · Score: 1

Eh, I don't believe him... there's no such thing as a dual Xenon system--it's probably just a single Xenon. However, in my spare time, I'm working on a 6.022E23 node Xenon Trioxide system... it'll be a blast!
Re:Mckenzie Cluster, faster, cheaper per TFlop by DaoudaW · 2003-08-23 16:06 · Score: 1

From the CITA McKenzie FAQ: At a total cost of CDN$900K (including tax), it is probably the best price performer in the world with a price:performance ratio of $0.75/Mflops.

The poster must have mistaken an M for a G.

This really make KASY0 look even better since it beat a competing claim for price/performance by a factor of six.

How are these booted? by cmason · 2003-08-23 03:49 · Score: 1

Am I missing something? They say:

KASY0 nodes are completely diskless; there isn't even a floppy. (from the FAQ)

So how are the nodes booted? Are there bioses out there that can netboot?

-c

--
"If you are an idealist it doesn't matter what you do or what goes on around you, because it isn't real anyway."-R.P.W.

Re:How are these booted? by digitalhermit · 2003-08-23 04:23 · Score: 1

Many ethernet cards have a socket for a programmable chip that allows netbooting. Pretty much all you need is the address of the server from where to retrieve the rest of the software. Usually the kernel is loaded via tftp then the rest of the os is NFS mounted. I don't know if this is how the article is doing it, but the netboot stuff is pretty common and easy to configure.
Re:How are these booted? by Anonymous Coward · 2003-08-23 04:26 · Score: 0

Perhaps bot BIOSes, but bootROMs. That's what they are for you know... either a ROM in a socket on the network cards or flashed into unused space in the BIOS chip.
Re:How are these booted? by afidel · 2003-08-23 05:46 · Score: 1

Yep, it's called IBA or Intel Boot Agent, it allows booting of a diskless system through PXE. It's actually where Paladium came from origionally. In order to pull a boot image over the network and be sure it was not tampered with on the wire through a man in the middle attack you need hardware crypto with a signed boot image. Basically every PC made in the last 4-5 years or so supports it (there are exceptions but they are usually consumer only oriented only PC's, corporate PC's almost all support it). It's how many enterprises install linux and windows if they don't use ghost.

--
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.

Re:Playstation2 at 5.5GFLOPS costs only $199 $40/G by Enonu · 2003-08-23 03:54 · Score: 1

The key word here is supercomputer. A PS2 is not a supercomputer.

This beat the PS/2 by Arker · 2003-08-23 03:57 · Score: 1

The previous price/performance champ was in fact a PS/2 cluster, mentioned here, but this AMD cluster is roughly three times the performance for the dollar. You can check the stats with different assumptions on their FAQ page, particularly the section labeled 'Is KASY0 really the first supercomputer under $100/GFLOPS?'

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.

New network architecture? by Anonymous Coward · 2003-08-23 04:00 · Score: 1, Interesting

Not to downplay the fantastic accomplishment, but there is nothing new about this network architecture. Not topologically, as Dietz has been claiming for years now. When did a mix of full and partial mesh suddenly become new?

Sweet cluster, though :D

Re:New network architecture? by mabhatter654 · 2003-08-23 14:55 · Score: 1

RTFA [i've always wanted to saythat] but seriously, they explained that the topology gains more performance from 100Mbs cards and switches than by simply getting "faster" stuff. They gain more performance from 3 normal 100Mbs cards than one would get from 1 Gbs card due to the time connections take to create and handle versus shareing the wire for multiple processes.
What's novel is that the same tech [crosslinking processors, etc.] that made the first crays so powerful can be emulated with cheap parts from your local store---that's huge!
Re:New network architecture? by Anonymous Coward · 2003-08-24 05:09 · Score: 0

While that's all well and good, it is not a new topology! I'm glad he's discovered the well-known fact that routing hops cause delay, and that you can bypass that by directly connecting nodes to their neighbors (well, just some).

Uh oh. Partial mesh.

Dehype, please, this is out of hand.

Re:Playstation2 at 5.5GFLOPS costs only $199 $40/G by Arker · 2003-08-23 04:02 · Score: 2, Informative

Gah feel free to mod the previous version of this comment into oblivion, I hit submit accidentally.

The numbers you're looking at are marketing numbers first off, and overly generous. Second you don't scale for free - you never get anything like 100 times the performance of a single box when you wire 100 together, for the same reason that you don't get twice the horsepower out of an engine twice the size.

The previous price/performance champ was in fact a PS/2 cluster, mentioned here, but this AMD cluster is roughly three times the performance for the dollar. You can check the stats with different assumptions on their FAQ page, particularly the section labeled 'Is KASY0 really the first supercomputer under $100/GFLOPS?'

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.

did you read the same article I did? by HanzoSan · 2003-08-23 04:09 · Score: 0

Did you read the article? Whats off topic?

Just because we have supercomputers does not mean we have the software or AI to take advantage of its power. Alot of people build junk just to say they built it, especially college students.

Mods, please mod the poster above as troll.

--
If you use Linux, please help development of Autopac

Re:did you read the same article I did? by imsabbel · 2003-08-23 04:15 · Score: 1

Yeah, and why do you think you need AI (?!?!?) to use its power?
There isnt enough computing power in the whole world for a few simulations some profs in the solid state faculty of my university would like to run.
its too bad if even a teraflop computer can only solve a 20^3 lattice...

--
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
Re:did you read the same article I did? by HanzoSan · 2003-08-23 12:47 · Score: 1

So 10 teraflops isnt enough?

--
If you use Linux, please help development of Autopac

There is Flop and Flop by Tiosman · 2003-08-23 04:14 · Score: 4, Insightful

It's not the first time that these folks in KY work around the definition of the acronym "Flop". A Flop is a floating point operation on 64 bits, not 32 bits. All entries in the Top500 used results with 64 bits HPL, nobody else in the world is running HPL on 32 bits. So claiming the moon on 32 bits is easy, useless for the sake of comparaison and almost unethical. I cannot believe that Dr Dietz do not know the difference by now.

The same machine would yield average results on 64 bits. Difficult to draw attention without headline numbers...

Re:There is Flop and Flop by Anonymous Coward · 2003-08-23 04:19 · Score: 0

Dietz is practically famous for overstating and overblowing this design. He wants every part of the project to be important, and the only thing that actually is, potentially, is the algorithm used to calculate the wiring.
Re:There is Flop and Flop by Anonymous Coward · 2003-08-23 05:16 · Score: 0

He uses a genetic algorithm to compute the wiring setup, so there's not even any novelty there. Coming up with some initial conditions and specifying constraints for a stock GA engine, hooray!

I've sat in on talks about the FNN and they weren't able to answer important questions such as the growth rate of switches/wires/ports necessary on machines as the number of machines grows to keep the "1-hop rule" true....
Re:There is Flop and Flop by Anonymous Coward · 2003-08-23 13:05 · Score: 0

EXCEPT... (from the faq page)

It really gets over 471 GFLOPS on a 32-bit version of HPL. Using an "untuned" 64/80-bit version, KASY0 gets a very respectable 187.3 GFLOPS. These aren't theoretical numbers, they are the real thing. The theoretical we-will-never-see-that numbers are 531 GFLOPS and 1.06 TFLOPS, respectively, for 64/80-bit and 32-bit floating point. ...
using 64/80-bit precision, KASY0's peak is better than $75/GFLOPS. On a single-precision version of HPL, KASY0's measured performance yields better than $84/GFLOPS.

the numbers still don't seem to add up for some reason though, but my brain has turned off for the day (and for a few weeks after stuffing 8 weeks of a diff equations class into 4 days , as i took the final exam as 100% of grade after slacking off all summer )

And next week ... by SmackCrackandPot · 2003-08-23 04:15 · Score: 1

... they're going to have the largest Quake LAN party ever!

Credit Card? by Hal+The+Computer · 2003-08-23 04:15 · Score: 1

Come on, 1 billion dollars on a credit card? Have you ever heard of a chargeback?

On a serious note I'm sure Paypal would be more (sic) secure and convienient.

--

int main(void){int x=01232;while(malloc(x));return x;}

Point of Trivia by jd · 2003-08-23 04:17 · Score: 1

Actually, many very early supercomputers were built into the basement/cellar for this very reason. Pack your computers as low as possible, and use the convection currents to carry the heat around the building.

(Those familiar with the University of Manchester's Department of Computation, in the UK, will understand what I mean. The architecture is designed around the computer room. Even after the truly massive lumps of iron were removed, it still wasn't until the mid 1990s that the building had a ground-floor entrance, as such an entrance would have limited the heating effect.)

--
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)

Actually... by jd · 2003-08-23 04:30 · Score: 1

There's a lot of OS essentials that could be moved easily into hardware. By using programmable gate arrays, or by just etching the kernel directly onto silicon, they should be able to reduce the energy requirements and thereby the actual cost.

Further, it would also accelerate the product enormously - Linux on a Chip would be blazingly fast, as it wouldn't take any processing power away from what it was running - thereby also reducing the cost per GFLOP.

--
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)

Yeah... by Cyno01 · 2003-08-23 04:49 · Score: 1

But if they cleaned it up they couldn't use it as a hammock.

--
"Sic Semper Tyrannosaurus Rex."

overclocking by snooo53 · 2003-08-23 05:03 · Score: 2, Insightful

Looking at the specs I'm curious if anyone thought of overclocking the machines to get an even bigger performance increase. It seems that with most Athlons you can get at least a good 100 mhz of extra speed, even with a stock cooler, by increasing the fsb/multiplier and not even touching the voltage. Even a modest increase like that would yield an extra 12.8GHz of power, dropping that price figure even further. Depending on what type of computing they're doing, increasing the fsb might have an even bigger effect than more GHz

Granted there might be some heat problems, but judging by their setup, I'm guessing the room is well-cooled.

--
The sending of this message pretty much inconveniences everyone involved.

Re:overclocking by Anonymous Coward · 2003-08-23 08:49 · Score: 0

My mom has a sewing machine. It has this neat peddle which acts like an accelerator. When you step down on it, the machine goes faster. One of these days I'm going to wire it up to my computer. Then I'll step down on it when I need more speed, and back off on it when I'm just surfin'.

Pardon me Cowboy by Stonent1 · 2003-08-23 05:12 · Score: 1

But the "FA" says $1000 per gflop not $100

Did you RTFA?

Re:Pardon me Cowboy by Anonymous Coward · 2003-08-23 05:16 · Score: 0

Sorry, crackpipe talking. It WAS $100.

No it doesn't by Anonymous Coward · 2003-08-23 05:24 · Score: 0

210A at 120Vac via the power law comes to 25.2kW, not kW/h (whatever the hell kW/h is supposed to stand for).

Users with IDs over 100 thousand shouldn't be allowed to post.

Re:No it doesn't by rusty0101 · 2003-08-23 05:38 · Score: 1

If you wish to calculate the cost of electricity you need to know the duration for which you are measuring. A kW/h is the measurement used by your elctric utility to determine how much they are going to charge you for the service they are providing. It is calculated by taking the power load, in this case 25.2kW and multiplying that by the hours that load is on the power grid (with 1 kW/h being the measure of a 1kW load being on the grid for one hour.)

24 of these a day, and [28-31] of days per month will provide you with the power load that this assembly will use. A standard 100 watt lightbulb will use 2.4 kW per day, and 74kW per 30 day month.

People with ID's under 6773437 shouldn't be allowed to post anonymously.

--
You never know...

Way to miss the whole point by pclminion · 2003-08-23 05:34 · Score: 1

That would be stupid. The entire point is, the nodes are so freaking cheap, that if you really want an extra 5% performance you just buy a few more nodes. Gee, what do I choose, buy a few more nodes, or spend two weeks overclocking all these finicky chips and trying to get them to run correctly?

Besides, nobody in their right mind would run a parallel program of any importance on a "rigged" setup like that.

Re:Way to miss the whole point by toddestan · 2003-08-23 05:59 · Score: 1

Not to mention if you can only overclock, say, 50% of them, would you run into problems with nodes running at different speeds?
Re:Way to miss the whole point by pclminion · 2003-08-23 06:04 · Score: 2, Interesting

It depends what sort of cluster it is. If you have a standard network of workstations, and you're running something like PVM or MPI, then each node can run at a different speed. In fact, they don't even have to be the same kind of nodes (you can have different platforms, say Solaris and Linux, both running in the same virtual parallel machine).Usually you will have to adjust your algorithms to account for nodes running at different speeds. But it doesn't make it impossible.
MOSIX is a parallel cluster operating system based on Linux that can run on nodes of different speeds. They all need to be the same platform, though -- you can't mix Sparc and Intel for example.
Other supercomputer applications are written specifically under the assumption that all nodes are the same speed, they are linked together a certain way, etc. It all depends on the application.
Re:Way to miss the whole point by snooo53 · 2003-08-23 12:25 · Score: 1

I have to disagree. Admittably, if you're building something on that scale you probably have money to burn. But looking at a cost/benefit analysis of it, I think the advantages to overclocking are enormous. With a modest amount of overclocking (~100mhz/machine for example), the risks are virtually nonexistant for that 5% increase. Maybe ten years down the road, the chips will start to suffer, but honestly who will care then? In ten years a $1500 home computer will most likely have this sort of processing power, making the whole setup obsolete anyway.
So I say, why not get the extra power while you can for free? Because a few years down the road it won't even be worth the electricity to keep this setup going.

--
The sending of this message pretty much inconveniences everyone involved.

Nigerian SCO scam by kaan · 2003-08-23 05:37 · Score: 1

DEAR SIR...

My name is Mumfasa Thumbutu, and I am writing to you from Nigeria. I work for SCO, and here in my country, we have many SCO license fees to collect. But due to our bad government we cannot put our money in the bank like normal people. See, everyone in Nigeria is corrupt, except for me and my friends (and SCO, too).

So we need your bank account number, social security number, mother's maiden name, date of birth, keys to your car, wallet, and your computer so that we can put the sum of ONE BILLION DOLLARS into your account.

This is not a scam, you see, because we are working for a reputable company (SCO) who has a lot of money.

SINCERELY,
MUMFASA THUMBUTU

(I tried posting this in all caps, but was thwarted by the lameness filter)

Not a slave when its your choice by vcbumg2 · 2003-08-23 05:54 · Score: 1

I work @ UK and the students used were on a pure voluntary basis

--

projects @ http://spectechnologies.net

In other news... by Anonymous Coward · 2003-08-23 05:55 · Score: 0

India breaks $9.99/GFLOPS barrier, using a team of veteran network engineers.

Imagine by Anonymous Coward · 2003-08-23 06:02 · Score: -1, Redundant

a Beowulf clust of those Beowulf clusters!

Spelled out, just for you by Anonymous Coward · 2003-08-23 06:16 · Score: 0

A kW*h is the unit you talk about. It makes no sense to talk about kW per hour (= kW/h), since kW is a unit of energy. It's like saying "kiloJoules per second per hour". Just senseless.

The rest of your post is just senseless. A light bulb uses an amount of kWh per day (or Joules, if you want to use SI units), not kW. And so on, and so forth.

Err by Anonymous Coward · 2003-08-23 06:19 · Score: 0

I meant kW is a unit of *power*, not energy.

Re:Err by Anonymous Coward · 2003-08-23 19:45 · Score: 0

Uhm perhaps you're trolling, but the other guy is right.
First of all Power == Energy
Second of all it is kW'h. kW is energy (or power) per second. You multiply it with hours (which are 3600 seconds) and you get energy (or power) since the /s goes away when multiplied with that unit. You pay for the amount of energy (or power) you use. You do not pay for the acceleration of your power usage (which'd be kW/h).
Re:Err by Anonymous Coward · 2003-08-23 19:48 · Score: 0

Argh seems I typod, supposed to be kW*h, obviously

What about $170K by Axe · 2003-08-23 06:20 · Score: 3, Funny

That they own to SCO, that damn commies? Did they at least aknowledge using stolen property?

What a shame. Freeloaders. They would never be able to achieve such performance if not for the fruits of labour of SCO .. eeeh.. lawers?

--
<^>_<(ô ô)>_<^>

$100/GFLOP by exp(pi*sqrt(163)) · 2003-08-23 06:23 · Score: 1

Er...you can do that with parts from ebay or craigslist without too much trouble.

--
Doesn't it make you feel good to know that our freedoms are protected by politicans, lawyers and journalists.

Smart by Anonymous Coward · 2003-08-23 07:00 · Score: 0

Nice end example. Indeed there are many factors to be considered in what architecture to use, These systems lower the cost and provide higher processing power for smaller, low budget, projects. I see great potential for these clusters to bridge the gap between grass root and high end projects.

Yes, beowulf clusters have amazing potential, bringing high end computing to the masses.

Electricity Cost by ModernGeek · 2003-08-23 07:11 · Score: 1

The site said that one year of power for the cluster would cost about as much as all the hardware cost. So I guess if you include power, it certainly doesn't beat the $100/GFLOPS barrier.

--
Sig: I stole this sig.

gain from cluster by h4x0r-3l337 · 2003-08-23 07:26 · Score: 1

It's interesting to note that the this cluster of 128, 2 GHz processors is only about 9 times faster than a single 1 GHz Itanium 2 processor at performing the PvRay benchmark.

Re:gain from cluster by Nynaeve · 2003-08-24 04:31 · Score: 1

It's interesting to note that the 1GHz Itanium 2 processor is 65 times more expensive (!) than the AMD Athlon 2600+
$7800 vs. $118
If you had a $40k research grant to build a supercomputer for less than $100 per GFLOP, which would you choose?
Re:gain from cluster by ahaveland · 2003-08-24 19:37 · Score: 1

Hi.. I'm the owner of the Povbench site... I believe the benchmark.pov file that the POVRay team chose to use was intended for single machines, not clusters. There is a good minute or so of processing time duplicated on all machines to calculate photon maps, and this information is not shared. If we subtract the time for this process, we approach something more accurate. I would really like to make a benchmark that doesn't spend so long precalculating photon maps and light buffers etc, so we can get a time just for pure distributed processing. However, it is a problem getting everybody to use the standard conditions, even if there is an option in the render menu to run benchmark! This only applies in the Windows GUI version. Other platforms have to make sure all the switches are standard. I hope this can be standardized across all platforms more easily in the future. Cheers, Andy.

university price calcuations are bogus by peter303 · 2003-08-23 09:42 · Score: 1

You have to include a people time, building overhead etc. A reserach grant may be billing $500 - $1000 a day for this. If this takes 50 man days to set up, then the cost is is another $50,000.

Re:university price calcuations are bogus by Anonymous Coward · 2003-08-24 05:52 · Score: 0

You have to include a people time, building overhead etc. A reserach grant may be billing $500 - $1000 a day for this. If this takes 50 man days to set up, then the cost is is another $50,000.

It is actually way worse than that. We've built several clusters (our latest one has 600 2.8GHz Xeon CPUs). Sure, "commodity of the shelf" hardware is cheap, but the quality is lower. It doesn't matter for a single machine, but with a couple of dozen boxes, components *WILL* break (even in the highest quality cluster).

Our new cluster is all rackmounted, and each node has a diagnostic LCD on the front. If a node goes down we just read what the problem is, call our hardware vendor, and within four hours that node is replaced. It takes about 5 minutes of our time, including the phone call. All fileservers and frontend nodes have backup power supplies and raid5 SCSI disk arrays.

Or... you could go the cheap PC way (as we've done before). Since you don't have any remote access card it takes a week before you find out a node is down. Then you have do spend an hour bugsearching the node, disassemble it and send the faulty part in. Five days later you get the part, put it in the node and get it online after another hour of work. 2-3 hours of a scientist's time is several hundred dollars wasted, not to mention that the total offline time is 1-2 weeks.

The important number isn't the cost per flop. It is the total amount of flops you ACTUALLY GET over the cluster lifetime, divided by the total cost of ownership, including your work.

I will never again built a cluster with more than ~20 nodes without using rackmounted nodes with remote access cards...

shelving by da2 · 2003-08-23 10:51 · Score: 1

that shelving they used looks almost identical to the kinda stuff i have in my bedroom, unfortunatly my shelves are not filled with a supercomputer (yet)

Re:You Can! that's the point by mabhatter654 · 2003-08-23 14:38 · Score: 1

They even have a tool on their site to help you wire it up! You can buy the hardware at any computer shop. Of course buying several hundred PC's at once and multiple switches takes a pretty heft credit limit! --Oh and you need to figure out how to power several hundred dual Athalon toaster ovens from your meger 200 amp service! But no one's stopping you. I think I'd be cool to do this to an office of PCs. After all an extra $50 per PC for a hobby project might get by the PHB.

Re:Braging rights! Nuf said! by mabhatter654 · 2003-08-23 14:43 · Score: 1

Any Uberslashdoters working on this?

we worship thee!

Slavery-ha, If I was available [unfortunately I'm in michigan] I'd be there for free too! Besides college students have fewer rights than slaves--every one knows that...

But STEVE JOBS told me by Anonymous Coward · 2003-08-23 16:38 · Score: 0

that the g5 was the fastest computer.

(for $39,500 you can buy only 10 g5!)

It's worse than that. by BiggerIsBetter · 2003-08-23 18:51 · Score: 1

The students are actually PAYING to be their University's slaves. I guess that makes it legal, if somewhat masochistic.

Side note: I wonder if my postgrad supervisor would make good dominatrix...

--
Forget thrust, drag, lift and weight. Airplanes fly because of money.

Re:Its about time - metaverselab, huh? by Anonymous Coward · 2003-08-23 21:15 · Score: 0

Okay now it's time to add in the licensing fee to Neal Stephenson for using 'metaverse'.

Maybe he'll go easy on them if they claim to be running finux...

128+4 by beggarstune · 2003-08-24 00:41 · Score: 1

What're the extra 4 for? They don't mention it. Hot spares? Command + control? Scheduling?

--
(S+C) x (B+F)/T = V

Re:128+4 by Nynaeve · 2003-08-24 04:39 · Score: 1

RTFAQ:
KASY0's configuration is:

* 128 + 4 "cold spare" PC nodes, each containing:
o One AMD Athlon XP 2600+ (the 2.075GHz version)
o One 512MB PC2700 DDR SDRAM
o BioStar M7VIT Pro motherboard
o Two Linksys LNE100TX NICs
o Codegen 6042L case with 400W power supply
* 18 BenQ SE0024 24-port Fast Ethernet switches
* 405 Cat5 Fast Ethernet cables
* RedHat Linux 9.0, modified Warewulf 1.11

I submitted this 2 months ago. by sundling · 2003-08-24 03:16 · Score: 1

I guess you have to pick a more interesting article on the same topic or know someone. :(

But they *can* be redundant! by Anonymous Coward · 2003-08-24 04:58 · Score: 0

OK, so they're not off-topic - they are likely to be "obvious to the skilled practitioner" :-) And yes, I'd mod myself down, but I'm posting it as AC...

Less heat because no video cards or disks by billstewart · 2003-08-24 05:21 · Score: 1

Remember that these things are pretty much just a CPU, RAM, and NICs - there aren't any video cards or disk drives (and as they comment, not even a floppy to boot from.) Your typical Gamer PC has a high-end video card, and some of those put out enough heat these days that they're coming with fans on them. So, yes, the CPU is cranking, but that's the only thing making much heat.

Now, there's the subtlety that they're doing SIMD stuff using the 3D-Now vector processing, which may run a bit hotter than regular random behaviour, but I don't know if it really does.

--

Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks

No windows, unfortunately by billstewart · 2003-08-24 05:26 · Score: 1

They did comment that the machine is located in the basement and doesn't have any windows, so it's dependent on the building's cooling system. Otherwise they could have taken care of much of the heat load by simply blowing the hot air outside, at least if they can tolerate Kentucky summer heat/humidity ambient conditions, which is a lot more efficient than actual airconditioning.

--

Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks

reported tuesday, july 29 - but incorrectly cited. by tfulton2 · 2003-08-24 12:38 · Score: 1

http://slashdot.org/comments.pl?sid=72589&cid=6558 445 Sorry for the sloppy reporting; posted in a followup the citation.

He is giving a talk on Oct 28 at 4:00 by tfulton2 · 2003-08-26 00:06 · Score: 1

...in the W. T. Young library on the U of Ky campus here in Lexington. Anyone in the area who wants to grill him should try it the old-fashioned way.

Slashdot Mirror

Supercomputer Breaks the $100/GFLOPS Barrier

281 comments