Merged into ISO C++ The functionality described on this page was merged into the mainline ISO C++ standard as of 11/2024; seethe data-parallel types (SIMD)(since C++26)

The SIMD library provides portable types for explicitly stating data-parallelism and structuring data for more efficient SIMD access.

An object of typesimd<T> behaves analogue to objects of typeT. But whileT stores and manipulates one value,simd<T> stores and manipulates multiple values (calledwidth but identified assize for consistency with the rest of the standard library; cf.simd_size).

Every operator and operation onsimd<T> actselement-wise (except forhorizontal operations, which are clearly marked as such). This simple rule expresses data-parallelism and will be used by the compiler to generate SIMD instructions and/or independent execution streams.

The width of the typessimd<T> andnative_simd<T> is determined by the implementation at compile-time. In contrast, the width of the typefixed_size_simd<T, N> is fixed by the developer to a certain size.

A recommended pattern for using a mix of different SIMD types with high efficiency usesnative_simd andrebind_simd:

#include <experimental/simd>namespace stdx= std::experimental; using floatv= stdx::native_simd<float>;using doublev= stdx::rebind_simd_t<double, floatv>;using intv= stdx::rebind_simd_t<int, floatv>;

This ensures that the set of types all have the same width and thus can be interconverted. A conversion with mismatching width is not defined because it would either drop values or have to invent values. For resizing operations, the SIMD library provides thesplit andconcat functions.

[edit]Main classes

[edit]ABI tags

[edit]Alignment tags

[edit]Where expression

[edit]Casts

[edit]Algorithms

[edit]Reduction

[edit]Mask reduction

[edit]Traits

[edit]Math functions

All functions in<cmath>, except for the special math functions, are overloaded forsimd.

[edit]Example

#include <experimental/simd>#include <iostream>#include <string_view>namespace stdx= std::experimental; void println(std::string_view name,autoconst& a){std::cout<< name<<": ";for(std::size_t i{}; i!=std::size(a);++i)std::cout<< a[i]<<' ';std::cout<<'\n';} template<class A>stdx::simd<int, A> my_abs(stdx::simd<int, A> x){    where(x<0, x)=-x;return x;} int main(){const stdx::native_simd<int> a=1;    println("a", a); const stdx::native_simd<int> b([](int i){return i-2;});    println("b", b); constauto c= a+ b;    println("c", c); constauto d= my_abs(c);    println("d", d); constauto e= d* d;    println("e", e); constauto inner_product= stdx::reduce(e);std::cout<<"inner product: "<< inner_product<<'\n'; const stdx::fixed_size_simd<longdouble,16> x([](int i){return i;});    println("x", x);    println("cos²(x) + sin²(x)", stdx::pow(stdx::cos(x),2)+ stdx::pow(stdx::sin(x),2));}

a: 1 1 1 1 b: -2 -1 0 1 c: -1 0 1 2 d: 1 0 1 2 e: 1 0 1 4 inner product: 6x: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 cos²(x) + sin²(x): 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

[edit]See also

[edit]External links

• In other languages

• 日本語
• 中文