Ilya Burylov

My hello world example, which I use to make initial build possible have only 3 remaining errors: ``` In file included from /std-simd/experimental/simd:59: /std-simd/experimental/bits/simd_builtin.h:1144:9: error: incomplete type 'std::experimental::parallelism_v2::_CommonImplX86' named in...

> Try this: > > ```diff > @@ -405,7 +405,12 @@ __is_neon_abi() > ⸱ > // }}} > // __make_dependent_t {{{ > -template using __make_dependent_t = _Up; > +template struct...

Basic arithmetic ops look working. Next broken thing is here: ```cpp // _S_signmask, _S_absmask{{{ template static inline constexpr _V _S_signmask = __xor(_V() + 1, _V() - 1); ``` clang does...

`stdx::abs` is now working for integers, but is returning zeros for FP for unknown reason. `stdx::fabs` works as expected with FP. `stdx::sin` has compilation issues

`_S_allbits` and `_S_signmask` are fine, but `_S_absmask` is not ```cpp __andnot(_S_signmask, _S_allbits); // does not work, I have no idea why auto a =_S_signmask; auto b = _S_allbits; __andnot(a, b);...

Last unresolved error for stdx::sin() to compile: ```cpp /std-simd/experimental/bits/simd_builtin.h:1663:12: error: call to '__and' is ambiguous return __and(__x._M_data, __y._M_data); ^~~~~ /std-simd/experimental/bits/simd.h:4569:14: note: in instantiation of function template specialization 'std::experimental::parallelism_v2::_SimdImplBuiltin::__bit_and' requested here...

Ok, my stdx::sin example works, and provides reasonable numbers. Got a number of similar warnings though: ```cpp In file included from /std-simd/experimental/simd:62: /std-simd/experimental/bits/simd_x86.h:2452:25: warning: '__builtin_is_constant_evaluated' will always evaluate to 'true'...

Tried to compile tests. Running via regular make quickly breaks with errors, thus I've applied direct approach to see wider picture for time-being. ```bash for testfile in *.cpp; do $CXX...

Then I'll continue trying to build it here. But I need some insightful help. Here is an error I get from tests ``` ./../experimental/bits/simd_builtin.h:2724:14: error: no matching function for call...

I am observing strange behavior with `long double` so far. Here ABI is deduced as `scalar`: ```cpp void foo(stdx::simd_mask a); ``` But that thing is deduced to `_VecBuiltin(16)` and it...