[simd] Add missing \libconcept to uses of same_as

[eisenwave]

This follows the precedent of https://eel.is/c++draft/simd.complex.access#3.2, which says

same_as<...> is modeled

We could probably also use is_same_v here instead, but that would create a local inconsistency with [simd.complex.access], and the three uses here are clearly wrong.

[hewillk]

There are nothing wrong with "If same_as<A, B> is true", we use a lots in <ranges>, see https://eel.is/c++draft/ranges#range.iota.view-10.1
The statement "If same_as<A, B> is model" is unnecessary because we are only checking the constraint here, and we are not asking about its semantics, which is much simpler.
I honestly don't see any improvement from is_same_v, except for being more verbose.

[jensmaurer]

Adding \libconcept is fine and expected.

Switching to "is modeled" is a potentially normative change (causing UB if the semantic implications of a concept are not met), whereas using the concept-id as a boolean expression is totally fine and only checks the "satisfied" part.

Please limit the editorial change to adding \libconcept.