Need help on improving performance of a small subfunction

I think it’s power x^p which blocks it unless you add @fastmath, although I’m not sure why with p::Int. (With a float, @fastmath (-0.1)^0.5 avoids the error.)

I don’t think the exponential usually SIMDs in Julia I’ve always been using Sleefpirates to get exp to SIMD, or lately, LoopVectorization