4 ms·

Comes to mind .. "The latest USB-C chargers are apparently more powerful than Apollo 11’s computer" https://www.theverge.com/tldr/2020/2/11/21133119/usb-c-ank

• mananaysiempre 3y ago
  I mean, with all due respect, so? An OG ATmega-based Arduino is significantly more powerful than the AGC. A garage opener from a decade ago can easily be more powerful (MIPS-wise) than the AGC. I wouldn’t be surprised if that holiday candy box that plays music when you open it were more powerful than the AGC. The reality of today’s MCUs is that you can get single-digit MIPS at the price of shipping, handling, and the plastic of the chip’s package. So you see those used where a handful of discrete logic would suffice, like before that you’d see discrete logic where a couple of relays and a handful funny-shaped mechanical actuators would suffice. It’s overkill in the sense that you don’t need nearly as much compute as you get, but it’s completely rational in that it’s genuinely the most cost-efficient approach. You can also (and thank you STMicroelectronics and before that Atmel for getting Microchip off their collective asses on that) get ridiculous amounts of I/O to drive with that compute. Gobs of GPIOs with individually configurable driving modes? No problem. Multiple DACs and ADCs? Sure. I2C and SPI? You bet. Two or three UARTs? Gotcha, would you like some CAN with that? In a 24-pin package so you can actually use more than a couple of these? Ahh now that’ll be four times the price, please and thank you. You know what’s also expensive? RAM and (to a lesser extent) ROM. The TRS-80 came with 4K bytes of RAM at minimum. The AGC had about the same computing power, 2K words of RAM and 36K words of ROM, and still the designers traded speed for program memory by writing a bytecode VM. The cheapest STM8S0, by contrast, comes with 256 bytes of RAM and no external memory interface; and I don’t even want to think about the obscure Chinese 6502 and 8051 clones you’ll find under the epoxy in toys and TV remotes. All that power and all you can fit in there is a glorified flowchart. But if you think about it as programmable logic, in the widest possible sense of the phrase, it makes perfect sense! Also (and here I could be wrong), my understanding is that the software on the AGC couldn’t actually plan the mission—it could do some specific things that you can’t reasonably to by remote, like landing, but most of the math was done by much chonkier programs on much beefier machines. It was a controller in the purest sense: it could control the hardware to achieve what you wanted, but you had to tell it precisely what, it didn’t figure it out for you. So if anything goes wrong and your comms are busted, you die. Don’t get me wrong, the AGC is bloody impressive. (Transparent native code / bytecode interworking anyone?.. None of those words were even in the dictionary back then.) But through equally impressive efforts of other people, its computing power was table stakes within a decade. Nowadays it’s a rounding error in your bill of materials. ETA: Ah, I see, the argument is that the Cortex-M0+ in question is superior in MIPS, RAM, and ROM. Yeah, that’s more reasonable. Low-end Cortex-Ms can also be very cheap these days[1], provided you can source them. [1] https://hackaday.com/2023/02/16/puya-py32-the-cheapest-flash-microcontroller-you-can-buy-is-actually-an-arm-cortex-m0/ https://hackaday.com/2023/02/16/puya-py32-the-cheapest-flash...
• yjftsjthsd-h 3y ago
  There's probably a commentary there about the compute being less important than it being attached to the most powerful rocket ever built.