Using -Ofast -mavx2 and a recent glibc, g++ vectorizes #include <cmath> void f(double*d){ d=(double*)__builtin_assume_aligned(d,256); for(int i=0;i<1024;++i) d[i]=std::exp(d[i]*std::log(2)); } However, if I write d[i]=std::exp2(d[i]) instead, it fails to vectorize (libmvec does not provide a vector version of exp2). It would be good, when checking if a standard function like exp2 has a vector version, to also check related, more canonical functions (exp in this case). (this could also be vaguely related to PR 81706)
Confirmed. glibc uses a different alg for exp2 than exp but for exp10 it simply does exp(x * log(10)) it seems. Probably requires some refactoring in vectorizable_simd_clone_call. Or we decide that with -funsafe-math-optimizations canonicalizing expN/logN to exp/log is worthwhile. The best thing would be to enhance glibc of course.