[Bug tree-optimization/127510] GCC vectorizes "std::for_each_n" loop taking on random access iterator with more than one pointers as uncounted loop

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127510

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
So this is an exit condition like

void foo (int *x, int sn, int sm, int n, int m)
{
  int i = sn;
  int j = sm;
  if (n != 0 || m != 0)
    do
      {
        x[i++] = 1;
        j++;
      }
    while ((i != n) & (i != m));
}

so there are N IVs that need to arrive at their exit value at the same
time for the loop to terminate.  This might or might not be something
to optimize separately - that is, rewrite in IVCANON (which requires
niter to be computable...).  Rewrite { a, +, CST }, { b, +, CST }
with same CST to { 0, +, CST } and replace uses by init + IV and then
version the loop on the condition it can terminate?  That is, the
exit condition is quite unfortunate - is this somehow controlled by
libstdc++ and can we fix it there?