Similar to cloud computing, there is another concept called “Flash Mob Computing”. As the name says, it refers to people bringing their computers to form an ad-hoc computer cluster trying to achieve supercomputer speeds. The idea as explained above is simple. People bring their laptops/plain old dekstops, connect them to a network, form an ad-hoc computer cluster and achieve massive computing speeds.
The first ever such event was called, “FlashMob 1” and was held at the University of San Francisco Gym and was organized by Slashdot. The peak rate achieved was a mind boggling 180 GigaFlops (for the noobs : Flop = Floating Point Operations per second). The best complete run could achieve a rate of only 77 GigaFlops. The cluster ran on Morphix Linux. That’s nothing compared to the IBM Roadrunner’s 1.456 PetaFlops, but it still is awesome. The 1.456 PetaFlops is infact 20,800 times the stable speed achieved at “FlashMob 1”.
FlashMob 1, held on April 3, 2004
Since then, the idea has kind of died. I couldn’t find evidence of any other such event. I understand that it won’t be a meeting for running some calculations for a big investment firm or for a scientific theory, but the concept of trying to create a supercomputer from notebooks and your very own notebook/desktop being a part of such an event is like awesome.
The IBM Roadrunner, the meanest machine on the planet
The famous list of the world’s Top 500th supercomputers managed by Hans Meuer, lists the 500th fastest supercomputer as the “BladeCenter HS21″ Cluster formed using Intel Xeon dual core 2.66 GHz processors at some financial services in the United Kingdom. It achieves a theoretical peak rate of 37.64 TeraFlops and the maximum LINPACK performance that has been achieved is 17.09 TeraFlops.
Coming to the point of holding another such event. According to Intel’s website, a 2.2 GHz Intel Core 2 Duo T7500 can achieve a rate of 17.6 GigaFlops under stable conditions. As this is the processor that I’m using on my notebook and it has become a fairly standard processor in modern notebooks, I’ll base my calculations on it.
(17.19 x 10^12)/(17.6 x 10^9) = 971.0222
972 such computers would do the job but let’s up it a little bit. If we have 1500 computers packing such a processor, we can achieve a maximum of 26.40 TeraFlops and that easily gets such a cluster into the Top 500 supercomputers list. In fact, at such a stable speed it would be the 203rd fastest computing machine on the planet.
The idea not only sounds awesome, it feels awesome as well. At least I’d feel powerful if a bunch of strangers who have collaborated over Facebook and Text Messaging, come together and create such a mad computing machine.
