On Fri, Oct 4, 2019 at 4:23 PM Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx> wrote: > How is it relevant whether the boot CPU is A5 or A7? These are bL > little cores that only implement NEON for feature parity with their bl > big counterparts, but CPU intensive tasks are scheduled on big cores, > where NEON performance is much better than scalar. Yea big-little might confuse things indeed. Though the performance difference between the NEON code and the scalar code is not that huge, and I suspect that big-little machines might benefit from unconditionally using the scalar code, given that sometimes they might wind up doing things on the little cores. Eric - what did you guys wind up doing on Android with the fast scalar implementation?