A Tcl article and an Erlang article - good morning!
I miss working with Erlang especially, but it's also certainly kind of a niche thing. Other languages are faster and have more effort being put into them.
Plenty of microcontrollers have a single-digit number of Cortex-M cores and memory/flash counted in the megabytes. It'll be decades until that market reaches the multi-gigabyte point, so why bother wasting a whole bunch of memory on 64-bit pointers?
I'm not quite sure why you'd want to run Erlang on it, but the hardware exists.
> I'm not quite sure why you'd want to run Erlang on it, but the hardware exists.
https://nerves-project.org/#features has a decent pitch for why. (Most of the features listed here aren't features of Nerves-the-Elixir-IoT-runtime-codebase per se, but rather benefits of Nerves-the-toolchain enabling you to easily build lean, embedded Erlang [on Linux] firmware images.)
> I'm not quite sure why you'd want to run Erlang on it, but the hardware exists.
Erlang is invented before IoT was a thing to facilitate distributed computing for telecommunication in a highly reliable manner. It makes perfect sense to adapt it for driving fleets of cheap IoT devices.
That does not mean ARM32 implementations and uses are stopping any time soon. Afaik arm hasn’t even obsoleted armv6, although Linux distributions are starting to drop it.
Either TBH, I imagined the main issue would be ram, even with psram. EQMX is used a lot for IOT and it'd be interesting seeing more heavy loads on the edge.
I miss working with Erlang especially, but it's also certainly kind of a niche thing. Other languages are faster and have more effort being put into them.
[1] https://www.grisp.org/blog/posts/2025-06-23-jit-arm32.1#why-...
Their existing hardware is aarch32. It really is that simple.
0. https://www.grisp.org/hardware
Plenty of microcontrollers have a single-digit number of Cortex-M cores and memory/flash counted in the megabytes. It'll be decades until that market reaches the multi-gigabyte point, so why bother wasting a whole bunch of memory on 64-bit pointers?
I'm not quite sure why you'd want to run Erlang on it, but the hardware exists.
https://nerves-project.org/#features has a decent pitch for why. (Most of the features listed here aren't features of Nerves-the-Elixir-IoT-runtime-codebase per se, but rather benefits of Nerves-the-toolchain enabling you to easily build lean, embedded Erlang [on Linux] firmware images.)
Erlang is invented before IoT was a thing to facilitate distributed computing for telecommunication in a highly reliable manner. It makes perfect sense to adapt it for driving fleets of cheap IoT devices.
It's the only ISA on Cortex-A32, but not sure if any mainstream chips were ever produced with that core.
(Depending on course if you want to get specific about Arm/Thumb/Thumb2, I lumped them all together above).
That said, if you're putting something like Erlang on a chip, aren't one likely to want the extra memory (and performance) of a slightly newer SoC.
https://www.cadence.com/content/dam/cadence-www/global/en_US...