@danbarber @aj @armbytes @deBaer @Kroc @dgar @nixCraft It was very significantly faster. I tested them. I know.
You needed to spend A LOT of money on an x86 machine to run x86 code faster than a relatively cheap Archimedes could run it in emulation.
It could also compute factorial 1000 in two seconds at a time when the Β£20,000 AI workstation on which I was employed to work took twenty minutes to run the same code.
@simon_brooke @danbarber @aj @armbytes @deBaer @Kroc @dgar @nixCraft
The ARM 2 was a nice CPU: clean well-thought-out instruction set, and what seemed like a generous bounty of 32-bit registers at the time (16 of them!), and yes, it was unusually fast compared to the x86 and 68xxx chips available around the same time.