Use a dedicated test instruction rather than a workaround

Author: rhpvorderman

src/sequali/_qcmodule.c

@@ -2464,13 +2464,7 @@ find_four_matchers_avx2(const uint8_t *sequence, size_t sequence_length,
         uint8_t index = NUCLEOTIDE_TO_INDEX[sequence[pos]];
         R = _mm256_and_si256(R, _mm256_loadu_si256((__m256i *)bitmask[index]));
 
-        __m256i check = _mm256_and_si256(R, found_mask);
-        /* Adding 0b01111111 (127) to any number higher than 0 sets the bit for
-           128. Movemask collects these bits. This way we can test if there is
-           a 1 across the entire 256-bit vector. */
-        int check_int =
-            _mm256_movemask_epi8(_mm256_adds_epu8(check, _mm256_set1_epi8(127)));
-        if (check_int) {
+        if (!_mm256_testz_si256(R, found_mask)) {
             bitmask_t Rray[4];
             _mm256_storeu_si256(((__m256i *)Rray), R);