Slashdot Mirror

yup. top500

If Intel says they can be used for anything... I say beowulf them all! We can show that the Top 500 (See #44) can be beaten! Hell, I'm trying to build one myself.

They could just "buy a Cray", except that SGI long ago stopped any additional development to the T3E line. This line was exactly like this - high speed Alpha CPUs (600 MHz on the -1200s) and incredibly fast, low latency interconnect. These machines still hold 24/50 of the top spots on the top500 list. Not bad for a computer designed in the mid-90s and last revised in early/mid-98. If they had allowed a continuation of the line (instead of promoting their SGI Origins instead), you would've been hearing Cray as opposed to DEC (err...Compaq). They're really just doing a good job of filling up the gap that SGI/Cray left wide open for them.

You do this so you don't have to actually set off a nuclear device to predict how it will perform. You must do a number of real nuclear tests to get baseline information. After that, you can do computer simulations of similar explosions. These simulations are VERY processor-intensive (like weather prediction) and require these large parallel systems to compute.

The U.S. does the same thing with massively parallel systems at Sandia National Labs and Lawrence Livermore National Lab. Check out the list of the top 500 supercomuter sites in the world -- http://www.top500.org/ -- to see who's doing this kind of thing.

Add special purpose processors and the efficiency and speed goes through the roof.

What's cool about Beowulf is that it lets non-US countries have access to supercomputers that they otherwise could not buy from the US. There are still significant restrictions in place that prevent many companies and governments outside of the US from buying high-end computers.

A /. article last year mentioned that Clinton had raised the limit on what's considered a supercomputer (so today's standard microprocessors can be sold), but it's still a problem.

Yes, Japanense companies (Fujitsu, Hitachi) make some very good supercomputers too. But my point is that places that want to have world-class supercomputers (at least the type that Beowulf offers) can now do so without needing to dance with US commerce restrictions to buy from SGI/CRAY, IBM, Compaq/Digital or HP.

For the Top 500 supercomputers in the world, see top500.org.

I wonder what a discount broker is doing with the twelfth biggest number-cruncher in the world.

Not true! While clusters and in particular Linux clusters are coming on strong, there are things that they just can't do.

The bandwidth of the interconnects on a Beowulf and worse still their latency are just not there yet to go head to head with a traditional supercomputer. They are getting very close, I admit, but haven't consistently beaten a Cray T3E-type interconnect yet.

Also, some codes/algorithms just don't lend themselves well to massively parallel implementations. They might be much happier on SMP-type machines or perhaps on vector machines.

Finally, some of the management issues for very large Linux clusters aren't fully resolved yet, but they have been in place on traditional supercomputers for quite some time.

As a result many institutions, including government, but also research sites and large financial institutions continue to buy Cray, SGI and SUN supercomputers, all of which aren't clusters. Just check the latest Top500 list, in particular the slides and statistics.

Remember, we're talking large systems here, and I would define this as more than 16 nodes and more than 32 CPUs minimum.

All the biggies are SMP nowadays. Look at the Top 500 Supercomputing Sites and you'll see what I mean. As for Cray, the first Cray company was bought by SGI, and the second went out of business, even though they had GHz single CPUs in about 1994 (exact info anyone?).

--

--

Good rebuttal but:

>> There are no commercially proven clustering >>technologies to provide High Availability for >>Linux."
> Can you say beowulf?

is false. Beowulf is a high performance clustering technologie. We still don't have a good high availability clustering solution AFAIK. There is work being done here but we aren't here yet.

But I am still waiting to see a NT machine on the top 500 list (the 500 more powerful supercomputers).

The Disc. Channel prog was good. It didn't pull punches when showing what they had. They said the computer (that could do the crack in under a second) was called a "Thinking Machine" it's actually a CM-5 (connection machine 5) made by Thinking Machines, Inc. Lotsa places have CM-5s, they're one of the most popular production supercomputers. Univ's like U. of Illinois (NCSA) and others have them availible to students.

They showed in the video a Cray T90-class supercomputer. Another popular one. These are nice systems (as far as supercomps go. they're just nice, not great.)

I assume that the NSA also has several of Cray's flagship models - the T3E-1200. Check out www.top500.org to see where I get my assumptions. The list is here.

The NSA has an affinity for very fast computers. They can use them to brute force just about anything.

Private companies have think tanks for coming up with math algorithms. Wolfram Research has some of them, they use them in Mathematica (a program). Mathematica has many secret algorithms for searching for prime stuff (numbers, factors, etc.).

There are other networks the NSA (presumably) uses to spy on people. One of them is rather obscure. Ask yourself this... "GPS uses like 24 satellites in polar orbit to cover the earth with a signal to tell you where you are. These are military (i.e. NSA) satelliites. They have the whole satellite to themselves. These are not little laptops in the sky - they're supercomputers. So what else do they put on the (several schoolbus sized) things???". Answer: Lots of goodies to make their jobs fun. Of course all of the things are top secret (even how GPS works). One of the things is a microwave camera. Ever use a cellphone in a building? Those signals go right through the walls (like they're not even there.) So does the light from these cameras. Ever had an x-ray done? You can see your bones. The freq. range they use is somewhere in between, so they can take pictures through walls, but get more than just bones on the "film". This is all well and good, but we can do better. Take three or more of these cameras and aim them at the same thing... What do you get? A 3-D image of an entire building whose contents show up through the walls.

Next time you're on the crapper, hemroids flairing, wave hi to the sky - we're watching.

The Disc. Channel prog was good. It didn't pull punches when showing what they had. They said the computer (that could do the crack in under a second) was called a "Thinking Machine" it's actually a CM-5 (connection machine 5) made by Thinking Machines, Inc. Lotsa places have CM-5s, they're one of the most popular production supercomputers. Univ's like U. of Illinois (NCSA) and others have them availible to students.

They showed in the video a Cray T90-class supercomputer. Another popular one. These are nice systems (as far as supercomps go. they're just nice, not great.)

I assume that the NSA also has several of Cray's flagship models - the T3E-1200. Check out www.top500.org to see where I get my assumptions. The list is here.

The NSA has an affinity for very fast computers. They can use them to brute force just about anything.

Private companies have think tanks for coming up with math algorithms. Wolfram Research has some of them, they use them in Mathematica (a program). Mathematica has many secret algorithms for searching for prime stuff (numbers, factors, etc.).

There are other networks the NSA (presumably) uses to spy on people. One of them is rather obscure. Ask yourself this... "GPS uses like 24 satellites in polar orbit to cover the earth with a signal to tell you where you are. These are military (i.e. NSA) satelliites. They have the whole satellite to themselves. These are not little laptops in the sky - they're supercomputers. So what else do they put on the (several schoolbus sized) things???". Answer: Lots of goodies to make their jobs fun. Of course all of the things are top secret (even how GPS works). One of the things is a microwave camera. Ever use a cellphone in a building? Those signals go right through the walls (like they're not even there.) So does the light from these cameras. Ever had an x-ray done? You can see your bones. The freq. range they use is somewhere in between, so they can take pictures through walls, but get more than just bones on the "film". This is all well and good, but we can do better. Take three or more of these cameras and aim them at the same thing... What do you get? A 3-D image of an entire building whose contents show up through the walls.

Next time you're on the crapper, hemroids flairing, wave hi to the sky - we're watching.

US energy has a relatively low cost of about $1/watt-year. Non-US energy costs are generally higher. For many energy efficiency is not a significant individual cost. Yet, consider the following...

CASE 1. Embedded High Performance Computing.

Sustainable PERFORMANCE/WATT can be a strong processor selection criteria for *embedded* processor selection including embedded high performance computing. Hence, vendors in the niche embedded high performance market tend to use PowerPC and DSP chips, not Alpha, Pentium or SPARC for dense compute solutions.

Vendor examples: CSPI, Mercury, Sky

Did folks notice that Black Lab Linux (see optional software) is working with MPI Software Technologies? Look close, MPI Software Technologies is actively supporting *embedded* MPI & MPI-RT.

CASE 2. Not Too Large Clusters (50 or so processors)

Also, notice the HPC Wire news article posted on the Top 500 site where...

"Several problems were caused by the power consumption and the heat in that small laboratory. The Paderborn people needed the fire brigade to pump cold air into the room. The power consumption was about 10 - 12 KWatt. Switching on only the power supplies of the nodes, the electric fuses switched off."

CASE 3. Blue Mountain...

6,144 processors -- 10,000 KW always online; 2,600 KW average usage.

SIXTEEN A/C Units plus FOUR 750-ton chillers plus TWO 10MW power stations (major UPS feature)

CASE 4. Aggregate World Energy Consumption (Green perspective)

Now consider 100,000,000 processors on desktops (far less than one per world capita)

100,000,000 * 30 watts-year * US$1/watt-year = US$3 Billion each year (without consideration for cooling costs, higher average world energy prices, CPU fan MTBF costs & such)

Hey, what's US$30 / year per CPU for a typical US home wallet??? Well maybe not much to many reading this... however, the energy costs can add up. Embedded system designers, large scale system installations, and environmentally sensitive people know what I'm talking about.

-- Think Global, Act Local --

US energy has a relatively low cost of about $1/watt-year. Non-US energy costs are generally higher. For many energy efficiency is not a significant individual cost. Yet, consider the following...

CASE 1. Embedded High Performance Computing.

Sustainable PERFORMANCE/WATT can be a strong processor selection criteria for *embedded* processor selection including embedded high performance computing. Hence, vendors in the niche embedded high performance market tend to use PowerPC and DSP chips, not Alpha, Pentium or SPARC for dense compute solutions.

Vendor examples: CSPI, Mercury, Sky

Did folks notice that Black Lab Linux (see optional software) is working with MPI Software Technologies? Look close, MPI Software Technologies is actively supporting *embedded* MPI & MPI-RT.

CASE 2. Not Too Large Clusters (50 or so processors)

Also, notice the HPC Wire news article posted on the Top 500 site where...

"Several problems were caused by the power consumption and the heat in that small laboratory. The Paderborn people needed the fire brigade to pump cold air into the room. The power consumption was about 10 - 12 KWatt. Switching on only the power supplies of the nodes, the electric fuses switched off."

CASE 3. Blue Mountain...

6,144 processors -- 10,000 KW always online; 2,600 KW average usage.

SIXTEEN A/C Units plus FOUR 750-ton chillers plus TWO 10MW power stations (major UPS feature)

CASE 4. Aggregate World Energy Consumption (Green perspective)

Now consider 100,000,000 processors on desktops (far less than one per world capita)

100,000,000 * 30 watts-year * US$1/watt-year = US$3 Billion each year (without consideration for cooling costs, higher average world energy prices, CPU fan MTBF costs & such)

Hey, what's US$30 / year per CPU for a typical US home wallet??? Well maybe not much to many reading this... however, the energy costs can add up. Embedded system designers, large scale system installations, and environmentally sensitive people know what I'm talking about.

-- Think Global, Act Local --

So if I understand correctly, this machine will be very useful for FERMI, if it is made to run the software they use over at top500, it would give worse results say than a ~300 nodes beowulf cluster?

maybe we'll see it in the TOP500 list? next update is 10-12 June 1999, hummm, too short...
--

I think loyal slashdotters everywhere should remind the keepers of that web site of the existence of the Beowulf class supercomputer, don't you?

---

Here is a quarterly survey of the top 500 most powerful systems in the world. FYI: Currently, the ASCI Red is at the top. Maybe we'll see Star Bridge Systems (SBS) up at number 1 soon.

whell all i can say is when are we going to have the computing power of some of the worlds fastest computers in a laptop. take a look at this on top500.org ,thier amazing. if thoes are the fastest computers, imagin how many pentiums and clones exist, and what if, just what if we installed pvmd on every computer in the world and each had a 1000BaseTX connection to each other and higher with the faster computers, what type of computing power would we have, only we can dream.

whell thats a thought ive had a lot, once i understand how to program for pvm ill take all my computers and all my roomates computers and link them :). i also had a "i rember when" conversation with a friend that was telling me when he was in the army the he was in the first team to use digital equptment :), and i sit hear and play with old AND and NOR gates, and think of how they even made a processor out of them. just my 2cents.