> > Are we really
> > expecting vendors to create and maintain a decent implementation of
> > <simd> for their DSPs from scratch?
>
> Yes. The basic implementation is where the Multiple in SIMD is just 1 where
> all operations are simply scalar. That suffices to meet the requirements of the
> header.

That's not what I would call a "decent implementation". That's
basically malicious compliance.

We technically allow these things so we don't go into the business of
making QoI mandatory and so these features don't have to be made
optional, but when all SIMD operations are scalar, it may as well not
exist. It would be a waste of time to implement SIMD like that in the
first place, rather than keeping the feature-test macro undefined and
expecting libraries to fall back onto some reasonable scalar behavior
themselves.

If you take a good look at what an implementation actually needs to
support to meet basic user expectations of quality, C++ really has
become too big for smaller vendors to roll on their own.