The same loop over vectors of uint8_t and uint16_t

Both functions are compiled with Clang or GCC at -O2 for arm64, using libc++ or libstdc++. How does the generated code for the two loops differ?

main.cpp
#include <cstddef>
#include <cstdint>
#include <vector>

void scale8(std::vector<std::uint8_t>& v, std::uint8_t k) {
    for (std::size_t i = 0; i < v.size(); ++i) v[i] *= k;
}

void scale16(std::vector<std::uint16_t>& v, std::uint16_t k) {
    for (std::size_t i = 0; i < v.size(); ++i) v[i] *= k;
}

40 of 44 in Compilers