Vectorizing Unicode conversions on real RISC-V hardware
camel-cdr.github.io
camel-cdr.github.io
Binary compatibility seems to be some kind of invisible glue that is often overlooked but holds software ecosystems together, e.g. people buying brand-new Steam Decks to run age-old Win32 exes.
I wonder what kind of unexpected developments will grow on a simple, stable and microarchitecture-agnostic ABI like RISC-V. A plethora of niche JIT compilers maybe?
Do you really need JIT if every targeted device has the same ISA?