It's often possible to avoid fringes via padding/overallocation. And if not, it may be possible to use unaligned loads/stores to handle the fringe in a single (

It's often possible to avoid fringes via padding/overallocation. And if not, it may be possible to use unaligned loads/stores to handle the fringe in a single (final) iteration: https://github.com/google/highway#strip-mining-loops https://github.com/google/highway#strip-mining-loops It is actually feasible to write vector-style code using SIMD instructions. Yes, the SIMD ISA is more complicated because of the various accumulated extensions, but this is what we currently have. And a bit larger code size (for one final loop iteration) doesn't seem to be a big deal.