c++ - What would be a good implementation of iota_n (missing algorithm from the STL)

Question

With C++11, the STL has now a std::iota function (see a reference). In contrast to std::fill_n, std::generate_n, there is no std::iota_n, however. What would be a good implementation for that? A direct loop (alternative 1) or delegation to std::generate_n with a simple lambda expression (alternative 2)?

Alternative 1)

template<class OutputIterator, class Size, class T>
OutputIterator iota_n(OutputIterator first, Size n, T value)
{
        while (n--)
                *first++ = value++;
        return first;
}

Alternative 2)

template<class OutputIterator, class Size, class T>
OutputIterator iota_n(OutputIterator first, Size n, T value)
{
        return std::generate_n(first, n, [&](){ return value++; });
}    

Would both alternatives generate equivalent code with optimizing compilers?

UPDATE: incorporated the excellent point of @Marc Mutz to also return the iterator at its destination point. This is also how std::generate_n got updated in C++11 compared to C++98.

Answer 1

As a random example, I compiled the following code with g++ -S -O2 -masm=intel (GCC 4.7.1, x86_32):

void fill_it_up(int n, int * p, int val)
{
    asm volatile("DEBUG1");
    iota_n(p, n, val);
    asm volatile("DEBUG2");
    iota_m(p, n, val);
    asm volatile("DEBUG3");
    for (int i = 0; i != n; ++i) { *p++ = val++; }
    asm volatile("DEBUG4");
}

Here iota_n is the first version and iota_m the second. The assembly is in all three cases this:

    test    edi, edi
    jle .L4
    mov edx, eax
    neg edx
    lea ebx, [esi+edx*4]
    mov edx, eax
    lea ebp, [edi+eax]
    .p2align 4,,7
    .p2align 3
.L9:
    lea ecx, [edx+1]
    cmp ecx, ebp
    mov DWORD PTR [ebx-4+ecx*4], edx
    mov edx, ecx
    jne .L9

With -O3, the three versions are also very similar, but a lot longer (using conditional moves and punpcklqdq and such like).