3 ms·
Some guys did ask chaGPT(3.5) to write a vectorized quicksort in x86_64 avx2 then avx512 assembly, then SPIR-V (GPU). It seems it is not that too bad for small
by sylware 4y ago
Some guys did ask chaGPT(3.5) to write a vectorized quicksort in x86_64 avx2 then avx512 assembly, then SPIR-V (GPU).
It seems it is not that too bad for small "well-known" algorithms. I am thinking high level language "ports" toward "human" assembly.