Modular Computing

[youtube=http://www.youtube.com/watch?v=ZBFoFYhC9B4]

This is the Illuminato X Machina, a “cellular” style computing system.  Each unit is a fully functioning computer with its own processor, storage and communications.  You can watch above as a change in the operating software is propagated across the grid. You can see the LEDs in the video going nuts, there are actually LEDs on the sides too. [Justin] described it to us as a personal fireworks show on your desk.  This system is fully open with the schematics and source code available on their site. You might recognize these guys too, we covered their Open Source GameBoy.

20 thoughts on “Modular Computing

  1. That totally reminds me of Stargate Atlantis and the replicators… My thought then is the same as now… how was it that an outside influence was able to inject code into such an advanced system so easily? I like this idea of cellular or distributed computing even with tiny robots each on their own but grouping up as a whole to accomplish a task… but I really hope we are able to secure this stuff before it actually makes it to a real product. *smile*

  2. I are not good speller.

    I am really interested in seeing the code – i.e. the communication protocol between chips and how the interface works such that devices can be connected in any orientation.

  3. I’ve always wanted to build a larger robot with modular computing like this. then have one chip designated as the brain which would update the other chips accordingly and send them sub commands.

    So say you have one chip for each leg, one chip per torso segment, and the brain chip. Then the brain chip can tell each leg to walk or make a pre programmed movement, while correcting itself from its sensor readings. If only I had a few hundred extra dollars and some propeller chips . . .

  4. “This system is fully open with the schematics and source code available on their site” I must be blind cause I haven’t found neither the schematics nor the source code :(

  5. Many years ago a similar concept existed, unfortunately for INMOS, the transputer concept never kicked in… maybe they were ahead of their time… Flash was something new at the time, and no one had it integrated into a MCU. Nevertheless, each transputer could download a program into his internal ram and execute it, each had 4 high speed serial connections and they were meant to communicate…. T805 was 32 bit @ 20MHz with FPU and comm links @5Mbps… quite impressive for the time…

  6. >I am really interested in seeing the code
    >– i.e. the communication protocol between
    >chips and how the interface works such that >devices can be connected in any orientation.

    We (Myself + other people on the course) had to write a similar thing to this but with 8051 boards with a pair of serial ports.. that was a good few years ago, and in 8051 assembler. Anyhow, this isn’t really all that hard to do if you can get your micro to do serial comms and can think up a basic protocol. Hardest bit will be writing the code to receive the new code and flash it without nuking itself. Atmel has an appnote on rewriting an avr’s flash during operation I think. Some chips have a section of flash reserved for bootloaders like this.. Google knows all.

  7. Chuck Moore’s multicore chips are much more interesting and combine 40 processors in a chip running at _very_ low power and are asynchronous (ie they run as fast as the silicon will allow). Current configuration has nine of these networked on a board to form a 360 cell computer. Details can be found at:

    http://www.colorforth.com/haypress.htm

    He is also developing other similar machines with 4 and 144 processors per chip. Details under the GreenArrays section of his website ( http://www.colorforth.com )

  8. I would turn an edge connector on some units up, some down, so you could 3D stack them, in addition to planar stacking, or perhaps use a modular carrier board and each unit would stack up or down to the carrier, so they could be removed on an individual basis for maintenance/repair.

  9. >>Imagine a style like this implemented in mobile robots. Throw in a random factor, and you might be able to observe some genetics at work.

    You could simulate this purely in software. Build in some ‘genetic rules’, even reproduction, and you really could observe genetics at work…

Leave a Reply to JonamCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.