4393. simd::unchecked_scatter_to is underconstrained

Section: 29.10.8.11 [simd.permute.memory] Status: New Submitter: Hewill Kang Opened: 2025-09-29 Last modified: 2025-10-03

Priority: Not Prioritized

View all issues with New status.

Discussion:

Both simd::unchecked_scatter_to and simd::partial_scatter_to are used to write a simd::vec into a range R.

However, they only require R to be contiguous_range and sized_range. Requiring R to be output_range is also necessary; otherwise, the constant_range cannot be written.

Proposed resolution:

This wording is relative to N5014.

1. Modify 29.10.3 [simd.syn] as indicated:

   namespace std::simd {
     […]
     template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
       requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
       constexpr void
         unchecked_scatter_to(const V& v, R&& out,
                              const I& indices, flags<Flags...> f = {});
     template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
       requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
       constexpr void
         unchecked_scatter_to(const V& v, R&& out, const typename I::mask_type& mask,
                              const I& indices, flags<Flags...> f = {});
   
     template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
       requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
       constexpr void
         partial_scatter_to(const V& v, R&& out,
                            const I& indices, flags<Flags...> f = {});
     template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
       requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
       constexpr void
         partial_scatter_to(const V& v, R&& out, const typename I::mask_type& mask,
                            const I& indices, flags<Flags...> f = {});
     […]
   }
   

2. Modify 29.10.8.11 [simd.permute.memory] as indicated:

   template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
     requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
     constexpr void unchecked_scatter_to(const V& v, R&& out, const I& indices,
                                         flags<Flags...> f = {});
   template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
     requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
     constexpr void unchecked_scatter_to(const V& v, R&& out, const typename I::mask_type& mask,
                                         const I& indices, flags<Flags...> f = {});
   

   -10- Let mask be typename I::mask_type(true) for the overload with no mask parameter.

   […]

   template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
     requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
     constexpr void
     partial_scatter_to(const V& v, R&& out, const I& indices, flags<Flags...> f = {});
   template<simd-vec-type V, ranges::contiguous_range R, simd-integral I, class... Flags>
     requires ranges::sized_range<R> && ranges::output_range<R, typename V::value_type>
     constexpr void partial_scatter_to(const V& v, R&& out, const typename I::mask_type& mask,
                                       const I& indices, flags<Flags...> f = {});
   

   -13- Let mask be typename I::mask_type(true) for the overload with no mask parameter.

   […]