amrex::simd::Vectorized< SIMD_WIDTH > Struct Template Reference

#include <AMReX_SIMD.H>

Inheritance diagram for amrex::simd::Vectorized< SIMD_WIDTH >:

Static Public Attributes
static constexpr int	simd_width = SIMD_WIDTH
	SIMD width in elements.

Detailed Description

template<int SIMD_WIDTH = native_simd_size_real>
struct amrex::simd::Vectorized< SIMD_WIDTH >

Mixin Helper Class

Use this class as a mixin (base) class to simplify writing functors that support/do not support ParallelForSIMD.

Example:

struct Compute : public Vectorized<>
{
    template<typename T_Real>
    AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
    void operator() (T_Real & AMREX_RESTRICT x) { ... }
};
 
// ... call site below
{
    Compute c;
 
    if constexpr (amrex::simd::is_vectorized<Compute>) {
        ParallelForSIMD<...>(np, c).
    } else {
        ParallelFor(np, c);
    }
}

Member Data Documentation

simd_width

template<int SIMD_WIDTH = native_simd_size_real>

constexpr int amrex::simd::Vectorized< SIMD_WIDTH >::simd_width = SIMD_WIDTH

staticconstexpr

SIMD width in elements.

---

The documentation for this struct was generated from the following file:

• /home/runner/work/amrex/amrex/Src/Base/AMReX_SIMD.H