inline xsimd::batch madd()

in gemmology.h [651:654]

• 4 lines of code
• 1 McCabe index (conditional complexity)

inline xsimd::batch<int32_t, Arch> madd(xsimd::batch<int16_t, Arch> x,
                                        xsimd::batch<int16_t, Arch> y) {
  return kernel::madd(x, y, Arch{});
}