Legos + Raspberry Pi ?

There seems to be a growing interest in the Raspberry Pi Foundation’s bare-bones, low cost (ca. $35) single-board Linux computer, the Raspberry Pi, designed originally for educational applications.  I noted earlier this month that the University of Cambridge Computer Laboratory was offering a free online course in operating system design, based on the Raspberry Pi.   This past week, Wired has published an article about another interesting development from the UK.

Though small, the little computer has respectable capabilities.  The first offering, the Model B board, mounts a 700 Mhz ARM CPU, a GPU, 256 MB of memory, audio, HDMI, and RCA video outputs. an Ethernet connection, and two USB ports; there is also a slot for an SD memory card.   Simon Cox, a professor of computational methods at the University of Southampton, originally bought a Raspberry Pi to use with his six-year-old son.  He then got the idea of using a number of the devices to build a supercomputer — with a case made out of Legos!

The resulting computer (the Wired article also has a set of photos) cost about £ 2,500 (a little over $ 4,000), not including networking equipment.  It has 64 Raspberry Pi computers, each with a 16 GB SD memory card, giving a total of 1 TB in addition to the on-board memory in each device.   The nodes communicate with each other using the Message Passing Interface standard, developed by the Argonne National Laboratory.  Professor Cox has a Web page with links to a number of resources, including instructions [PDF] for building your very own supercomputer.  You can even install a FORTRAN compiler.

The setup developed by Prof. Cox and associates is not ideal in some ways; for example, it requires 64 power supplies, one for each Raspberry Pi.  But it seems like an excellent tool for teaching students about massively parallel computing.

Top 500: Sequoia is Number One

Since 1993, the TOP500 project has been publishing a semi-annual list of the 500 most powerful computer systems in the world, as a barometer of trends and accomplishments in high-performance computing.   The systems are ranked based on their speed in floating-point operations per second (FLOP/s), measured on the LINPACK benchmark, which involves the solution of a dense system of linear equations.

The latest version of the list has just been released, in conjunction with the 2012 International Supercomputing Conference, currently being held in Hamburg, Germany.  The top system this time is the Sequoia system at the Lawrence Livermore National Laboratory, which clocked in at over 16 petaflops (16 × 10¹⁵ flops):

For the first time since November 2009, a United States supercomputer sits atop the TOP500 list of the world’s top supercomputers. Named Sequoia, the IBM BlueGene/Q system installed at the Department of Energy’s Lawrence Livermore National Laboratory achieved an impressive 16.32 petaflop/s on the Linpack benchmark using 1,572,864 cores.

The Japanese K-RIKEN system, ranked number 1 in the November 2011 Top-500 list, is now ranked second.  Ranked third is the Mira system at the Argonne National Laboratory, an IBM BlueGene/Q system with 786,432 processing cores, running at 8.15 petaflops.  The Chinese Tianhe-1A system, ranked second in November 2011 with 2.57 petaflops, is now ranked number 5.  The total capacity of the entire list is now 123.4 petaflops, compared with 74.2 in November.

As has been true for some time, the distribution of operating systems used is rather different from that in the desktop computing market:

OS Family	Number	% of Capacity
Linux	462	92.4
Unix	24	4.8
BSD-based	1	0.2
Windows	2	0.4
Mixed	11	2.2

Microsoft’s dominance of the desktop OS market clearly does not cut much ice in this area.

You can see the complete list here.