optimizations: SIMD / NEON

[ZLangJIT]

what is the current status for jit translation-to SIMD/NEON instructions in relation to the RISC-V instruction set ?

[LekKit]

The RISC-V Vector extension i think is not easily translatable to SIMD/NEON because it is a lot more advanced. AVX512 on the other hand is somewhat useful, but not many CPUs exist with it (I certainly don't have one).

Vector extension may be emulated via scalar loops, and if we write a C scalar implementation the compiler might be able to vectorize a few places.