Extend #1172 approach to avx512

Author: serge-sans-paille

include/xsimd/arch/xsimd_avx512bw.hpp

@@ -330,6 +330,41 @@ namespace xsimd
             return _mm512_cmpgt_epu8_mask(bool_val, _mm512_setzero_si512());
         }
 
+        template <class A, class T, class = typename std::enable_if<batch_bool<T, A>::size == 32, void>::type>
+        XSIMD_INLINE batch_bool<T, A> load_unaligned(bool const* mem, batch_bool<T, A>, requires_arch<avx512bw>) noexcept
+        {
+            __m256i bpack = _mm256_loadu_si256((__m256i const*)mem);
+            return { _mm512_sub_epi16(_mm512_set1_epi8(0), _mm512_cvtepu8_epi16(bpack)) };
+        }
+
+        template <class A, class T, class = typename std::enable_if<batch_bool<T, A>::size == 32, void>::type>
+        XSIMD_INLINE batch_bool<T, A> load_aligned(bool const* mem, batch_bool<T, A>, requires_arch<avx512bw>) noexcept
+        {
+            __m256i bpack = _mm256_load_si256((__m256i const*)mem);
+            return { _mm512_sub_epi16(_mm512_set1_epi8(0), _mm512_cvtepu8_epi16(bpack)) };
+        }
+
+        template <class A, class T, class = typename std::enable_if<batch_bool<T, A>::size == 16, void>::type>
+        XSIMD_INLINE batch_bool<T, A> load_unaligned(bool const* mem, batch_bool<T, A>, requires_arch<avx512bw>) noexcept
+        {
+            __m128i bpack = _mm_loadu_si128((__m128i const*)mem);
+            return { _mm512_sub_epi16(_mm512_set1_epi8(0), _mm512_cvtepu8_epi32(bpack)) };
+        }
+
+        template <class A, class T, class = typename std::enable_if<batch_bool<T, A>::size == 16, void>::type>
+        XSIMD_INLINE batch_bool<T, A> load_aligned(bool const* mem, batch_bool<T, A>, requires_arch<avx512bw>) noexcept
+        {
+            __m128i bpack = _mm_load_si128((__m128i const*)mem);
+            return { _mm512_sub_epi16(_mm512_set1_epi8(0), _mm512_cvtepu8_epi32(bpack)) };
+        }
+
+        template <class A, class T, class = typename std::enable_if<batch_bool<T, A>::size == 8, void>::type>
+        XSIMD_INLINE batch_bool<T, A> load_unaligned(bool const* mem, batch_bool<T, A>, requires_arch<avx512bw>) noexcept
+        {
+            __m128i bpack = _mm_loadl_epi64 ((__m128i const*)mem);
+            return { _mm512_sub_epi16(_mm512_set1_epi8(0), _mm512_cvtepu8_epi64(bpack)) };
+        }
+
         // max
         template <class A, class T, class = typename std::enable_if<std::is_integral<T>::value, void>::type>
         XSIMD_INLINE batch<T, A> max(batch<T, A> const& self, batch<T, A> const& other, requires_arch<avx512bw>) noexcept