Slashdot Mirror
Canada to Launch Countrywide Virtual SuperComputer
LadyCatra writes "A serious shortage of world-class computing power in Canada prompted University of Alberta scientists to create the next best thing -- a countrywide, virtual supercomputer.
On Nov. 4, thousands of computers from research centres across the country will be strung together by a U of A effort to create the most powerful computer in this country.
The full story is here"
Why didn't they just make a client program for distributed computing so the entire country/world could help out?
sig.
This seems like a really good idea, I don't really understand why more places don't do this. I mean most of us work in offices where the computer power is amazing and largely untapped.
I think what this really needs is to be make easier for the mainstream, so anyone could do it. Perhaps bundle the tools (programming and deployment) with mainstream operating systems?
It's just an idea, my NeXT had Zilla (it's version of this) years ago - seems a shame that this hasn't caught on more widely. So come on Apple - let's see it, put it in the Darwin project and put a nice UI on it in Mac OS X.
My school, in conjunction with the Université de Sherbrooke (mostly the U de S) are setting up a world-class beowulf cluster for general scientific work. A physics professor at my University, who also happens to be a world class astronomer (Dr. Lorne Nelson) has a research grant that he is using to help with the funding for this cluster.
So, why is this news? Is there some new technology they are using?
Do we? Idle time means the CPU is using less power, and generating less heat. I suppose that theoretically you are also making your processor transistors life slightly shorter, although there are probably arguments that a constant 50% CPU utilisation is not a bad thing because it will be more likely to maintain a constant temperature...
In any case, multiplied up by many millions of installed PCs, using that idle time means increasing energy consumption by a not insignificant amount. We need to use less energy, not more! Indeed, saying that idle time is "buring(sic)" away is quite the opposite of the truth.
Conversion Rate Optimisation French / English consultant
Not all CPUs power down when being idle. Most OSs has an idle task, burning away computer power in an endless loop.
When usage is 50%, the CPU is probably not turned off at all, since turning on and off clock trees (and getting the PLLs to sync) take time.
Since most home computers will not power down, we can use that potential computer power to save energy by not running super computers elsewhere.
Fair enough, good point. Can you confirm that the idle task in question makes the CPU heat up as much, and uses as much energy as a floating point operation continuously looping? I have a hunch that it doesn't...
Conversion Rate Optimisation French / English consultant
Saving up for a "real" supercomputer is a pipe dream. Supercomputers cost several million dollars a year in upkeep, and that's the killer. You might easily get grants to allow a project to use 'x' dollars worth of computing, but nobody is going to approve a capital grant that requires millions each year.
When the University of Toronto did purchase a Cray in the mid-eighties, there was a massive fight. Many felt that the resources to support the Cray were sucking money desperately needed everywhere else. (although, boy, we in meteorology a happy bunch...)
While lower profile and somewhat more painful to use, this is far more practical solution for the realities of academic computing today.
"Exactly what are you selling?"
I'd like to suggest something like the JavaVM, i.e. a standard virtual machine, from which you buy and sell basic ops, i.e. a byte-code instruction.
The biggest problem will probably be that you will not make any real money from letting your CPU be used. Perhaps a good idea would be to let a university supply you with internet access in exchange for CPU time. They usually have quite alot of bandwith.
As many of the other posters have pointed out, this work isn't necessarily new, but it is news.
There are other tools out there which do this: Legion, Avaki, Sun Grid Engine, Globus, to name a few but the goal is to create a network of (mostly) supercomputers which doesn't require a lot of reconfiguration at each site. What differentiates this work from many other approaches is that it is transparent to the system administrator.
For those who ask "why can't you just do something let seti@home" the answer is that not all problems in science and business can be easily decomposed into small chunks. Bandwidth requirements and latency may also be a problem. A lot of scientific programmers have to worry about communications much more than about processing power (although this tradeoff has been seesawing backwards and forwards with new advances in both technologies).
There's a worldwide effort through both business and academia to create a number of good, interoperating frameworks for doing this sort of transient, virtualised supercomputer.
Have a look at the Global Grid Forum (which is becoming the focus for Grid computing standards) for more information.
The University of Alberta has over a dozen clusters. Their central computing facility (CNS) has two clusters, Physics has three or more, CS at least one, Chemistry has seven clusters (0.5 THz total cycles), MechEng at least one, EE at least one, ...
The U of A (U of Eh?) also participates in MACI (www.maci.ca) and houses three SGI Origin computers, and is involved with the WestGrid project (www.westgrid.ca).
Prof. Schaeffer's point isn't that we don't have "computrons", but that research is increasingly using simulations (see Jaeger's work) and other computational methods, and computational resources are becoming increasingly overloaded as budgets are not growing as quickly as research advances.
--altadel