3 ms·

Why didn't they use an open source CPU? Eg: OpenSPARC

• avmich 11y ago
  May be they wanted good speed, freedom from particularly complicated FPGAs, and not to have to order chips as ASICs?.. I'd like to know too; this could be a good thing to have.
• xobs 11y ago
  Aside from the fact that we're familiar with Freescale and ARM, many other chips we "could have used" are unobtanium. While it's true that there is a source-level implementation, I can't find any T1 or T2 parts available for purchase. It's the same reason why we went with an A9 instead of an A15 or a 64-bit chip: You just can't buy them unless you're a big company. And if a small two-person company can't buy them, how can we claim it's open source hardware if you can't buy them either? The other possibility would be to fab a chip ourselves, but that's a whole other order of magnitude in terms of cost and complexity, and the result isn't that great in terms of speed and available peripherals. Plus, when you fab a chip like this a lot of the hardware blocks are IP provided by the chip foundry, e.g. flash controllers and DRAM cells, and those are always closed-source. It just moves the whole thing one turtle down.
• renox 11y ago
  Because you can't buy (1) an OpenSPARC? 1:for a reasonable amount of money, paying for a custom ASIC is quite expensive.