[PATCH 00/11] C++26 [simd]

[Matthias Kretz]

Tomasz Kaminski [Thursday, 23 October 2025, 13:05:13 CEST]:
> >  However, I'm not really happy with how the code reads now, because that
> > "is constprop" includes is_constant_evaluated isn't that obvious in the
> > name.
> > 
> > How about __is_value_known(...) instead of __is_constprop(...)?
> 
> What about __is_constknow? I do not care much about internal helper name,

Well, I read it a lot. So if it trips me up I need to change it. :)

> but when I see __is_value_know, I assume that answer is true, because the
> value is always known at runtime.

Interesting, even after you point that out I don't read it like that. In my 
unit tests I have a make_value_unknown(x) function that hides the value from 
the optimizer. I'm probably thinking too much from the viewpoint of the 
optimizer. :)

__is_const_known(...) then.

-- 
──────────────────────────────────────────────────────────────────────────
 Dr. Matthias Kretz                           https://mattkretz.github.io
 GSI Helmholtz Center for Heavy Ion Research               https://gsi.de
 std::simd
──────────────────────────────────────────────────────────────────────────