Hi, On 2022-06-28 11:17:42 +0700, John Naylor wrote: > On Mon, Jun 27, 2022 at 10:23 PM Hannu Krosing <han...@google.com> wrote: > > > > > Another thought: for non-x86 platforms, the SIMD nodes degenerate to > > > "simple loop", and looping over up to 32 elements is not great > > > (although possibly okay). We could do binary search, but that has bad > > > branch prediction. > > > > I am not sure that for relevant non-x86 platforms SIMD / vector > > instructions would not be used (though it would be a good idea to > > verify) > > By that logic, we can also dispense with intrinsics on x86 because the > compiler will autovectorize there too (if I understand your claim > correctly). I'm not quite convinced of that in this case.
Last time I checked (maybe a year ago?) none of the popular compilers could autovectorize that code pattern. Greetings, Andres Freund
