Slashdot Mirror
Supercomputers To Move To Specialization?
lucasw writes "The Japan Earth Simulator outperformed a computer at Los Alamos (previously the world's fastest) by a factor of three while using fewer, more specialized processors and advanced interconnect technology. This spawned multiple government reports that many suspected would ask for more funding in the U.S. for custom supercomputer architectures and less emphasis on clustering commodity hardware. One report released yesterday suggests a balanced approach."
Ignoring size, how does the cost of a cluster of fewer, highly specialized computers (with special interconnects, etc.) compare with that of a cluster of more, less specialized computers?
Teraflops per dollar is important, let's not forget that.
How does one go about bench marking a super computer specialized to do a certain task versus cheap computers in a cluster. Now we need to spend more money to develop specialized super computers even though the case scenerio presented in japan might not hold true to other applications? Seems a little too soon to start making recommendations
Skynet had 60 Teraflops IIRC and they're talking about 100!
Let's hope this isn't tied into Nukes somehow. Wait a sec, a massive virus has already spread disabling millions of computers!
RUN HIDE! THE END IS UPON US!!!!!!!
----
Go canucks, habs, and sens!
The Japan Earth Simulator outperformed a computer at Los Alamos (previously the world's fastest) by a factor of three while using fewer, more specialized processors...
What is the difference between processor designed to simulate earthquakes (et al) and an ordinary, off-the-shelf processor? I mean - so they optomized floating point operations. Is that it?
To make laws that man cannot, and will not obey, serves to bring all law into contempt.
--E.C. Stanton
What if you care only about integer operations?
You tell me how "whilst" differs from "while," and I'll stop calling you a pretentious jackass.
If you're going to have a supercomputer do one thing, of course specialize it. An Earth simulation surely has a set number of formulae whose calculations are to be optimized as much as possible, even to the hardware level.
But if you want a versitile, general-purpose supercomputer, why not go with the clustering solution?
The two studies resulted, in part, from NEC Corp.'s May 2002 announcement of the Earth Simulator, a custom-built supercomputer that delivers 35.8 teraflops. That system packed five times the performance of the fastest U.S. supercomputer at that time...
"The Earth Simulator created a tremendous amount of interest in high-performance computing and was a sign the U.S. may have been slipping behind what others were doing," said Jack Dongarra...
Graham said researchers should not overreact to NEC Corp.'s Earth Simulator that blindsided many in the high-performance computing community eighteen months ago by delivering a custom-built system five to seven times more powerful than the more off-the-shelf clusters developed in the U.S.
I don't mean to draw a crude analogy here, but I really can't help but read this and be reminded of the space race.
It took Sputnik to kickstart our spacemindedness; I for one consider it sad that a "tremendous amount of interest" -- and the funding that comes with it -- in high-performance computing seems only to have arisen/regenerated with the influence of competitive international politics. Are we really so hardly advanced that our respective national egos are still the driving force behind enthusiasm, financial or otherwise, in certain areas of science?
The coolest voice ever.
Is there a way to really compare the speed of a supercomputer and commodity hardware? If anyone could give either a quick explanation or a link to the relationships between bogomips teraflops MHz and the whole lot I would be very much appreciative.
-Silmarildur
Specialized hardware (almost) always outperforms commodity stuff.
I use custom designed amplifiers because they work better for my application. I could buy off-the-shelf stuff (~$500~$10,000 range), but that won't be exactly what I want. I use custom software too... know why? Because it's designed specifically for the job. That same software shouldn't really be used for other fields of research, neither should my amplifiers. The thing about this stuff is that it takes a lot of time to maintain (plus initial development). That means grad students, postdocs, and technicians who may spend over 90% of their time just keeping systems in working order and/or adding features. The benefits of customized hardware/software, in this instance, is worth the headaches associated with it.
All of my optics is commodity stuff (some is rare/exotic, but it's still basically black-box purchasing). I don't have the facilities to make coated optics, nor do I need anything that specialized, so... I just buy it.
When I was in telecom, we used Oracle and Solaris and Apache. It worked, and the cost of developing the same functionality in-house was ridiculously high (plus we'd never get to designing our products that sit on top of it).
Eventually, it always comes down to a comparison between the cost (man hours, equipment, etc) of custom building and of integrating stuff from OEMs.
So, the question our labs need to answer is, does clustered COTS hardware get the job done? Supplementary to that, is it cost-effective to buy/design it in light of the previous answer?
In any field where you are pushing the limits of technology, you have to make such trade-offs. Personally, I don't care who has the absolute fastest supercomputer (measured in flops, factoring-time, whatever)... what really counts is, who does the best research with the supercomputers.
Down with Saudi Arabia!!!
Specialized systems are almost always going to outperform generalized systems when you're dealing with similar levels of technology (for instance, specialized abacasuses vs. a generalized Cray T3E).
;))
The great thing about generalized systems is you can use them to explore new areas, then design a specialized system to take advantage of specific optimizations the generalized one can't support.
I'm glad for the report suggesting a "balanced approach". I can't imagine forsaking one type of system for the other, as each has its place. (Uhoh... generalized systems have a "place"? Does that mean they're specialized at being generalized? Oh, the irony!
bytesmythe
Hypocrisy is the resin that holds the plywood of society together.
-- Scott Meyer
I assume that hard-coding trig functions into the tertiary processors would be advantagious for this. I know it violates the spirit of RISC in general-pupose computing, but for such a large scale system with so many processors it coould be advantagious.
Do HP's Saturn or other such special-purpose processors have hard-coded higher-level functions?
You can't judge a book by the way it wears its hair.
Cray is back and getting back into the government contract game. Suprisingly, they are doing it just as the DOD is realizing that they need specialized hardware like they used to when Cray was one of their best suppliers. Look for little ol Cray to be back in the black real quick, and pick up a few shares now.
"Curiosity killed the cat, but for a while I was a suspect."- Steven Wright
Good question there man.
I am also wondering, which should I get? I mean, with Doom III on its way, to get decent frames should I go specialized supercomputer, or a linux beowulf cluster?
The main area in which we saw benefit was switching from the Portland Group Fortran Compiler to the Intel Fortran Compiler, which cut the timestep (simulation time/real time) nearly in half.
Every cluster in the department is assembled from commodity x86 components. Groups here have been moving from proprietary Unix architectures to Linux/x86 systems and clusters. Our group started out on RS/6000s, then moved to SPARC, and is now moving to x86. In terms of price/performance there really is no comparison.
As for TCO, the lifetimes of clusters here are relatively short, one or two years at the most. Thus a high initial outlay cannot be set by lower cost of operation.
Definitely a really huge super-computer would be neat to have but honestly are they putting the ones we already have to good use?
From what I've heard [anecdotally] computers like the earth simulator go vastly under utilized for the most part.
So given that most nations [including the US] have budget problems specially concerning education couldn't people think of better uses for money?
And before anyone throws a "it's the technology of it" argument my way, I'd like to add that if anything I'd rather have the money spent on researching how to make high performance low power processors [and memory/etc] instead. E.g. an Athlon XP 2Ghz that runs at 15W would be wicked more impressive than a 50,000 processor super computer that runs a highly efficient idle loop 99% of the time.
Tom
Someday, I'll have a real sig.
Stop being silly. The cooling requirements of an Athlon based massively powerful supercomputer would eat up the savings from using standard parts.
Seriously, though - I would guess, actually, that if one were to build a supercomputer from a "desktop" processor, the PPC970 (aka G5) chips would be a good choice. They have a solid vector unit, are RISCier, have a wider bus, and a better pipeline design. Plus IBM's fabrication capabilities are excellent - which helps in reliability and upgradability.
Frankly, I don't want the fastest computer chips on the desktop to be designed by a company in another country (even if Intel makes them outside of the US) and I would rather that the cutting edge, be cut here, in my native country.
Good lord, why? Is it just national/istic pride? I see that as something to be outgrown with respect to driving, receiving, and appreciating scientific discoveries and technological advancements. Honestly, if Japan were to come out with, say, the first mass-produced DNA computer, I wouldn't be the slightest bit bitter, or reluctant to take advantage of it. I regularly praise other countries for doing things the U.S. hasn't.
German physicists were primarily responsible for breakthroughs in their field in the 19th and early 20th centuries, and during that period there was quite a bit of resentment from American politicians and scientists whose feelings boiled down to nothing more than "We should have gotten there first." I won't argue that fierce competition has been beneficial to mankind at large (we've seen it in the computer industry, after all) but I don't think I'm wrong in wanting the motivation to be something a little less self-centered, political, immature. An idealistic vision? Hardly. It's not too much of an expectation for us to evolve beyond petty glare-throwing.
The coolest voice ever.
If you'd like to see what these people are up to for yourself, here is a link to their website. Lots of performance data, lists of projects, etc.
There seems to be an impression in some comments that this machine has some sort of special design that's only applicable to climate modeling problems. In fact, this is a vector-based supercomputer, applicable to any problem where you need to perform vector operations (i.e., operating on large arrays of numbers in parallel).
Certain numerical operations can be performed blindingly fast on these types of machines. Each arithmetic processor on this machine has 72 vector registers, each of which can hold 256 elements. Then you can perform operations on all 256 elements of 1 or more registers simultaneously! If the algorithm can keep the vector units fed, they will scream.
Since keeping data flowing to the processors is critical to speed, the high-speed interconnects (~12GB/s) are a must for any problem that is not completely localized. It's all about matching the problem to the hardware. There may well be problems for which a commodity cluster just can't get the job done like this can. Remember that each node of a cluster consumes power, produces heat, and takes up space. The raw cost of hardware is not the only consideration.
Earth Simulator is impressive in its own reguard, but in no way is the majority of clustering apps going toward these 'specialized' systems. Governments, research labs, etc. want powerful computers that are dirt cheap. Los Alamos's ASCI Q (Installment 1, the Alpha servers) cost over $100,000,000 to build, while their Pink cluster cost about $6,000,000 in hardware. On paper, Pink and ASCI Q are both around 10 teraflops. ASCI Q runs Quadrics on 64-bit 66MHz PCI, Pink is getting an ugprade to Myrinet Lanai 10 on PCI-X (From Lanai 9 on 64/66PCI). Not only that, but Pink runs the open-source, 100% GPL'd Clustermatic software and can be booted in a matter of seconds rather than hours like ASCI Q.
The fact is, systems like ASCI Q and the Earth Simulator just aren't practical. They may look great on paper, but there's not much that they can do that can't be done on x86. Given the choice between paying over a hundred million for a proprietary cluster that might not even be all that reliable (*cough*Q*cough*) and requires expensive software and maintenance contracts, we see companies like Linux Networx offering high-power clusters on common hardware and free software that are a fraction as expensive.
As far as reliability goes, don't get suckered into thinking that proprietary and expensive mean quality. Q's failure rate is almost as high as my old Windows '98 machine hahaha. With the exception of a few missing chillers, Pink seems relatively healthy with only a few minor failures.
If CRAY and NEC want to get into a pissing contest in specs, that's fine. If they offer something that Intel can't, more power to them. Otherwise, the five organizations in the world that own their systems can be proud that they have the most powerful computer on paper for a year or two before someone builds a cheaper x86 cluster that matches or out-performs them.
That is the whole point.
I have the feeling the DOE (nuclear weapon simulation etc) simulation program is not going anywhere near as well as it was sold.
Massive commodity clusters boast big numbers but they do not boast great useful throughput of USEFUL RESULTS. (also with massive clusters
you have to be able to deal with inevitable hardware failures).
You have a certain fluid problem---there is a certain speed of sound, and a certain physical geometry. What you want to do is to be able to simulate the real thing at ever smaller grid-sizes, that is, with greater numerical approximations to the physical fields.
Ideally, if your problem were embarassingly parallel and clusterizable, then you could put any number of grid points on each CPU and crunch away. You want more grid points? buy more CPUs.
The problem is that in actual physics the length scale of 'interaction' per time step does NOT go down---remember, speed of sound is constant as is physical geometry---imagine for instance the uh, radiative driven implosion of a certain unspecified dense material in spherical or cylinderical geometry into one unspecified not-dense material.
So when you scale-up in the scientifically useful sense---and not the computer nerd sense---then a problem which used to be solvable efficiently on clusters NO LONGER IS SO. There is just too much communication, and this is driven by physical reality.
It is not 'OK' to just say "change your code". The codes are developed with mathematical methods and based on experimental data gleaned over literally decades at great expense.
Programming for these is not easy---but it is quite a bit easier for the large vector old-skool cray type machines than the clusters, where the human has to do almost all the scutwork (e.g. MPI).
The problem is actually more severe with the DOE fluids problem---there are fundamental mathematical issues in the nearly inviscid flow (singular perturbation theory baby) which have not yet been resolved. And they appear at smaller and smaller grid sizes.
This requires rapid development of models and validation at the physically important resolutions and you can't do this with a cluster.
I have no inside information whatsoever but I smell that the sudden DOE and DOD interest in back-to-the-future retrosupercomputing is because of some major failures in the recent cluster efforts.
There is also a direct trade-off between more general purpose systems and systems custom tailored to a task. Good examples are Deep Blue and Blue Gene. Both of these systems are designed with a particular task in mind (i.e. chess and protein folding) and therefor are able to leverage knowledge about the problem space to constrain the kind of hardware, the particular low-level instructions and the information flow within the system while achieving signifigantly greater performance on a small class of problems. I work with clusters that are used in scientific communities that have various researchers working on various problems. In these cases, the questions are about basic applicability of a particular problem to a particular architecture. For example a cluster with high-speed interconnects made of good COTS hardware will allow a user with a very granular problem to effectively use the cluster and it will also allow a user who needs the high speed interconnect because the problem space demands a high degree of internal communication. But the first researcher might also be able to make use of a grid of (for instance) many more computers with a total lower cost because (s)he doesn't need the high speed interconnect. The Earth Simulator gains a lot of performance (on a class of problems) because of the underlying vector processor architecture. Given the right internal bus it is conceivable that adding vector processor daughter boards to the next generation of COTS clusters could achieve similar results--but, of course, only for problem spaces that make efficient use of such processors and aren't bottlenecked by the communication requirements.
Real answers are always more complicated. For example: the equations needed for nuclear simulation will probably require dedicated hardware (as the need for protein folding has lead to Blue Gene) to achieve the results that the Pentagon needs. But for many super computing tasks, the flexibility of COTS clusters will still be compelling, especially for areas where the algorithms are not yet fully developed (e.g. brain simulation). An interesting keynote at OLS 2003 argued that (some of) the problems are not going to be the local computing power but the need to move large quantities of data between research labs across the world and combine computational systems using the 'grid.' (For a down home examples of problems that have been successfully tackled through course granular distribution just look at SETI@Home and Distributed.Net. So its not just the flops anymore...
Would you rather they simulate weapons or resume detonation testing of new designs?? The fact is the US has a VERY large and ever aging supply of weapons, most of the cycle time so far from the ASCI projects has gone towards stewardship of the existing crop of weapons, making sure that the stockpiles are safe and also that they will be effective(if god forbid they should be needed). Also, reduced consumption is the only thing that will reduce our environmental "problems". Personally I think anyone who thinks the US has much of an environmental problem needs to get out of LA/New York/whichever big city they live in. I have spent a lot of time enjoying the national parks of this great country and I can tell you that there are a lot of pristine wilderness areas and a lot of generally green land here (in fact the US landmass is one of the least densly populated non-desert areas in the world).
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.