Rp2040 runs unmodified PalmOS 5.2.8
twitter.com
twitter.com
FWIW, you can replace twitter.com with nitter.net to see it (depending on why you don't want to visit twitter)
"I ended up writing my own kernel"
If you ever need a definition of system software engineering "tour de force", that is it.
Which of the various corpses of Palm assets/acqusitions did this stuff end up in, IP ownership wise?
Both interesting projects to be sure, but to me (and maybe the author) there's more satisfaction in running unmodified binary software on a readily available microcontroller.
So weird how Palm got stuck in a halfway state between architectures like that. I suppose it's not too dissimilar to where Apple was mid 90s with the 68k->PowerPC transition in MacOS.
I know it's fast compared to palm hardware but still.. microcontrollers tend to have io bottlenecks. Nice work.
The Amiga 1000 and Amiga 500 ran on a 7.09 (PAL) or 7.16 (NTSC) 68000 starting out with 256 KB (not MB) ram for the 1000, 512 KB for the 500. With this they ran a pre-emptive multitasking single-user OS which was partly ROM-based, partly loaded from floppy. These were responsive systems as long as they did not need to pull data from floppy drives.
So I imagine it was already pretty snappy on those devices. I think the RP2040 might be a downgrade in comparison...the clock speed at least, is much lower.