On Sat, Jun 03, 2023 at 10:53:04AM +0200, Mason wrote: > >> Sure, some builtins are officially supported. > >> Those are meant to be used directly by users. > >> > >> Then there are many builtins which are implementation detail for some > >> other API that users should use instead. > >> That includes e.g. builtins used under the hood for <*intrin.h> > >> implementation - users should use the intrinsics from those headers, > >> that is documented interface which is supported by multiple compilers, > >> or builtins used under the hood inside of libstdc++ headers (again, > >> users should use standard C++ APIs which are supported by multiple > >> compilers instead of the builtins directly) etc. > >> E.g. between GCC 3.4 and current trunk 62 __builtin_ia32_* builtins > >> which were implementation details of the x86 intrinsic headers > >> have been removed as the intrinsics got implemented some other way > >> (e.g. using generic vectors etc.). > > > > Does it matter whether or not those builtins are documented ? It seems > > like most of the __builtin_ia32_ builtins are explicitly documented in > > the manual, despite the fact that these seem like those you're referring > > to as being potentially removable at will whenever - and I see no > > indication in the documentation that they are implementation details/may > > be removed at any time for any reason. > > Hello Gabriel, > > As far as I understand, Jakub is merely saying: > If a builtin is NOT documented, then it is NOT part of the API. > > https://gcc.gnu.org/onlinedocs/gcc.pdf > *builtin_ia32_addcarry* appears nowhere in the manual, > thus it is NOT documented, thus it is NOT part of the API. Roughly, except that some builtins are mistakenly documented when they shouldn't be. I think e.g. none of the __builtin_ia32_* builtins are actually supported. The documentation even mentions some __builtin_ia32_* builtins which were removed years ago. E.g. __builtin_ia32_paddw128 but lots of similar ones in GCC 5 (2015). Jakub
