Try explicit impl

Author: AntoinePrv

include/xsimd/arch/xsimd_neon.hpp

@@ -767,7 +767,14 @@ namespace xsimd
          * add *
          *******/
 
-        WRAP_BINARY_INT(vaddq_u8, vaddq_s8, vaddq_u16, vaddq_s16, vaddq_u32, vaddq_s32, vaddq_u64, vaddq_s64, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_u8, uint8x16_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_s8, int8x16_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_u16, uint16x8_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_s16, int16x8_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_u32, uint32x4_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_s32, int32x4_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_u64, uint64x2_t, detail::identity_return_type)
+        WRAP_BINARY_IMPL(vaddq_s64, int64x2_t, detail::identity_return_type)
         WRAP_BINARY_FLOAT(vaddq_f32, detail::identity_return_type)
 
         template <class A, class T, detail::enable_neon_type_t<T> = 0>