5 ms·
It's indefinitely scalable exactly in the sense that a machine on this architecture can be made bigger as long as you can keep providing power, cooling, real es
by livcomp 4y ago
It's indefinitely scalable exactly in the sense that a machine on this architecture can be made bigger as long as you can keep providing power, cooling, real estate, and tiles. There is no head node, no unique tile addresses, no global communications, no global clocking, and no pretense of deterministic execution. In that sense it's quite different than traditional cellular automata. Figuring out how to do useful work in this architecture is the research challenge.
- gtsnexp 4y agoWhat aspect(s) in the premise of this interesting/exotic system suggests that useful work may be done by/with it?
- svnt 4y agoShifting contexts this much tends to shift existing intractable problems into new classes of tractable (and still intractable) problems. Answering this will require a lot of people with differing domain expertise discovering it and trying it out. This is a potential (I’m inclined to say inevitable) revolution, once all the pieces are in place.
- livcomp 4y agoYay thanks this!
- livcomp 4y agoThe most basic aspect is: Natural life does useful work, suggesting that artificial life might also. I'm like if you want a spreadsheet, fine, use a von Neumann machine. But if you want do inherently robust system control, that has a chance of doing something sensible even in situations that were neither programmed in nor trained upon, what you want is an overprovisioned system that is intrinsically aware of its deployment in space, and is constantly repairing and rebuilding itself.. and this video is another baby step on that road.
- gtsnexp 4y agoWould it be fair to say that the next challenge would be to develop a new kind of programming language architecture/concept that would take advantage of a t2-tile system? Otherwise, what is the strictest bottleneck?
- livcomp 4y agoWe've got the ulam programming language custom tailored for the MFM, and the SPLAT spatial programming language built on top of ulam. I expect we'll want more languages or language features as we scale up, but we need to earn our way to them with design experience. I'd love to build a T3 tile, probably FPGA-based, using lessons learned, that could provide perhaps 10x or 100x the average event rate of the T2s. And get them out into researchers and hackers in quantity.
- jacquesm 4y agoAny chance of re-purposing PiZeros for this? That might be a softer approach that can be tried relatively cheaply.
- livcomp 4y agoMaybe! One main challenge is the (at least) six-way local intertile communications. The T2s use BeagleBone Greens, which have two PRUs that I slice three ways each to do packet transfers. The RP2040 anyway has two 'PIO' instances that seem similar. But I'm unsure that any redo at that scale would improve delivered performance that much. If I had it to do over I think I'd've tried to put an ethernet router chip on each tile and basically do backplane ethernet between tiles. But I dream of LVDS serdes between tiles with low-level packet stuff handled in FPGA fabric..
- jacquesm 4y agoAre you aware of Jecel de Assumpção? If not you really should hit him up it sounds like the two of you would have a lot to talk about.