@steeph I struggle to see how at least my design would benefit from SIMD...
If I say that its microcode is lookuptable-branching, I don't think that'd scale beyond operating a byte-at-a-time. But existing RAM designs can supply those sequential bytes at a rapid rate!