Sometimes you get the feeling that no one really has any idea what’s going on.When everything is put into the vernacular you sometimes forget how amazingly specialized and complex the world you inhabit really is. But don’t you worry: There’s so little that you understand!
There are supremely smart people working very hard on almost everything! Think about that! How comforting!
Take this typical Wikipedia sentence, for instance:
“The entire BOINC network averages about 5.3 petaFLOPS as of July 25, 2011.”
Isn’t it blessedly abstract? What could it possibly mean?
An entire sentence of the English language that–to some experts–makes total sense. But this splendid sentence can also function as a keyhole through which the layman can glimpse an entire world of super-computing, alien searching and crude Wikipedia humor. So!
Let’s talk BOINC.
We’ll start with the subject of the sentence, which takes you most of the way: the BOINC network. BOINC stands for Berkeley Open Infrastructure for Network Computing. It is a network set up to use available computing power from personal computers—like the one I’m typing on now and perhaps even the one you’re using to read this—to use extra computing power for science.
You simply download the open source BOINC software, enter an email and run the program. While your computer sits there, its computing power is used to process highly complex computer models for things like climate models, a three-dimensional model of the Milky Way, or the attempt to break historical Enigma codes.
The best part of this whole thing, apart from the wonderful name, is that the BOINC technology and network was developed to help search for extraterrestrial life, by using a network of volunteer personal computers to sift through data collected by the Arecibo Radio Telescope looking for flashes of signals from across the universe.
Other projects have muscled in on this computing power, but thankfully almost 10 percent is still combing the skies for Afternoon Rush Hour DJs from the great beyond. In the last month, 22.07 percent of this computing credit has been put toward finding the largest prime number.
So we now know what the BOINC network is. And we know that it averages about 5.3 petaFLOPS as of July 25, 2011.
What is a petaFLOP?
A petaFLOP is a measure of computer speed. Like all tests—EPA gas mileage tests, a football player’s 40 yrd dash time, the SATs—it’s debatable how much the LINPACK-developed FLOP test tells scientists about real-world application. As Wikipedia describes it “the results from the LINPACK benchmark are compared by computer nerds in white coats, similar to the way teenage boys compare dick sizes,” which is a fairly compelling argument that unedited Wikipedia articles tell you a lot more than any edited encyclopedia ever could.
LINPACK was originally developed way back in the late 1970s, for measuring their puny supercomputers. The LINPACK benchmark is a measure of how many “floating point operations” (FLOP) a computer can run in a second (S). We’ll skip the in-depth on what exactly a floating point operation is; all you really need to know is that the computers are solving dense systems of linear equations, and the LINPACK benchmark is how we measure that, in the number of floating operations done per second. First they were measured in millions of FLOPS, can you imagine where we are now?
Just last month, the Japanese company Fujitsu burst through the 10 petaFLOP barrier aboard the mysteriously named “K Computer.” The “K Computer” will sit atop the coveted Top500 spot as the world’s most powerful supercomputer with an impressive 10.51 quadrillion floating point operations per second.
So the BOINC network averages about 5.3 quadrillion floating operations per second. A bunch of personal computers are hooked together going half as fast as the fastest computer ever made, looking for aliens and the highest prime number.
What a world!