Optimize 'construct at end' loops in vector

[mvels]

Summary

This change adds local 'end' and 'pos' variables for the main loop inmstead of using the ConstructTransaction variables directly.

We observed that not all vector initialization and resize operations got properly vectorized, i.e., (partially) unrolled into XMM stores for floats.

For example, vector<int32_t> v(n, 1) gets vectorized, but vector<float> v(n, 1). It looks like the compiler assumes the state is leaked / aliased in the latter case (unclear how/why for float, but not for int32), and because of this fails to see vectorization optimization?

See https://gcc.godbolt.org/z/UWhiie

By using a local __new_end_ (fixed), and local __pos (copied into tx.pos_ per iteration), we offer the compiler a clean loop for unrolling.

A demonstration can be seen in the isolated logic in https://gcc.godbolt.org/z/KoCNWv

The com

libcxx/include/vector

 //  Precondition:  __n > 0
 //  Precondition:  size() + __n <= capacity()
 //  Postcondition:  size() == size() + __n
 template <class _Tp, class _Allocator>
 void
 vector<_Tp, _Allocator>::__construct_at_end(size_type __n)
 {
     _ConstructTransaction __tx(*this, __n);
-    for (; __tx.__pos_ != __tx.__new_end_; ++__tx.__pos_) {
-      __alloc_traits::construct(this->__alloc(), _VSTD::__to_address(__tx.__pos_));
+    const_pointer __new_end = __tx.__new_end_;
+    for (pointer __pos = __tx.__pos_; __pos != __new_end; ++__pos, __tx.__pos_ = __pos) {
+        __alloc_traits::construct(this->__alloc(), _VSTD::__to_address(__pos));
     }
 }
 
 //  Copy constructs __n objects starting at __end_ from __x
 //  throws if construction throws
 //  Precondition:  __n > 0
 //  Precondition:  size() + __n <= capacity()
 //  Postcondition:  size() == old size() + __n
 //  Postcondition:  [i] == __x for all i in [size() - __n, __n)
 template <class _Tp, class _Allocator>
 inline
 void
 vector<_Tp, _Allocator>::__construct_at_end(size_type __n, const_reference __x)
 {
     _ConstructTransaction __tx(*this, __n);
-    for (; __tx.__pos_ != __tx.__new_end_; ++__tx.__pos_) {
-        __alloc_traits::construct(this->__alloc(), _VSTD::__to_address(__tx.__pos_), __x);
+    const_pointer __new_end = __tx.__new_end_;
+    for (pointer __pos = __tx.__pos_; __pos != __new_end; ++__pos, __tx.__pos_ = __pos) {
+        __alloc_traits::construct(this->__alloc(), _VSTD::__to_address(__pos), __x);
     }
 }
 
 template <class _Tp, class _Allocator>
 template <class _ForwardIterator>
 typename enable_if
 <
     __is_cpp17_forward_iterator<_ForwardIterator>::value,

 void
 vector<_Tp, _Allocator>::__move_range(pointer __from_s, pointer __from_e, pointer __to)
 {
     pointer __old_last = this->__end_;
     difference_type __n = __old_last - __to;
     {
       pointer __i = __from_s + __n;
       _ConstructTransaction __tx(*this, __from_e - __i);
-      for (; __i < __from_e; ++__i, ++__tx.__pos_) {
+      for (pointer __pos = __tx.__pos_; __i < __from_e;
+           ++__i, ++__pos, __tx.__pos_ = __pos) {
           __alloc_traits::construct(this->__alloc(),
-                                    _VSTD::__to_address(__tx.__pos_),
+                                    _VSTD::__to_address(__pos),
                                     _VSTD::move(*__i));
       }
     }
     _VSTD::move_backward(__from_s, __from_s + __n, __old_last);
 }
 
 template <class _Tp, class _Allocator>
 typename vector<_Tp, _Allocator>::iterator