generalization of reduce_mul implementation

[serge-sans-paille]

No description provided.

[serge-sans-paille]

Note to self: we should provide dedicated reduction for or, and and xor operator, and do it in a generic way if possible. RISCV supports those reductions.

[JohanMabille]

Technically, this means the next release should be a major one. On the other hand, since reduce_add was already available, I wonder if this function was really used.

[serge-sans-paille]

On Fri, Aug 29, 2025 at 08:45:48AM -0700, Johan Mabille wrote: @JohanMabille approved this pull request. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ In [1]include/xsimd/arch/common/xsimd_common_arithmetic.hpp: > @@ -139,20 +139,6 @@ namespace xsimd return fma(x, y, select(mask, neg(z), z)); } - // hadd - template <class A, class T, class /*=typename std::enable_if<std::is_integral<T>::value, void>::type*/> - XSIMD_INLINE T hadd(batch<T, A> const& self, requires_arch<common>) noexcept - { - alignas(A::alignment()) T buffer[batch<T, A>::size]; - self.store_aligned(buffer); - T res = 0; - for (T val : buffer) - { - res += val; - } - return res; - }

well, it's not part of the standard API, is it?

[JohanMabille]

Ah indeed, so we're good