add avx512 support for Lower/Upper function

This commit is contained in:
yaqi-zhao 2022-06-02 16:17:13 -04:00
parent 8c40f05d4c
commit 3929baf9ff

View File

@ -29,31 +29,61 @@ private:
{
const auto flip_case_mask = 'A' ^ 'a';
#if defined(__AVX512F__) && defined(__AVX512BW__) /// check if avx512 instructions are compiled
if (isArchSupported(TargetArch::AVX512BW))
{
/// check if cpu support avx512 dynamically, haveAVX512BW contains check of haveAVX512F
const auto byte_avx512 = sizeof(__m512i);
const auto src_end_avx = src_end - (src_end - src) % byte_avx512;
if (src < src_end_avx) {
const auto v_not_case_lower_bound = _mm512_set1_epi8(not_case_lower_bound - 1);
const auto v_not_case_upper_bound = _mm512_set1_epi8(not_case_upper_bound + 1);
const auto v_flip_case_mask = _mm512_set1_epi8(flip_case_mask);
for (; src < src_end_avx; src += byte_avx512, dst += byte_avx512) {
const auto chars = _mm512_loadu_si512(reinterpret_cast<const __m512i *>(src));
const auto is_not_case
= _mm512_and_si512(_mm512_movm_epi8(_mm512_cmpgt_epi8_mask(chars, v_not_case_lower_bound)),
_mm512_movm_epi8(_mm512_cmplt_epi8_mask(chars, v_not_case_upper_bound)));
const auto xor_mask = _mm512_and_si512(v_flip_case_mask, is_not_case);
const auto cased_chars = _mm512_xor_si512(chars, xor_mask);
_mm512_storeu_si512(reinterpret_cast<__m512i *>(dst), cased_chars);
}
}
}
#endif
#ifdef __SSE2__
const auto bytes_sse = sizeof(__m128i);
const auto * src_end_sse = src_end - (src_end - src) % bytes_sse;
const auto v_not_case_lower_bound = _mm_set1_epi8(not_case_lower_bound - 1);
const auto v_not_case_upper_bound = _mm_set1_epi8(not_case_upper_bound + 1);
const auto v_flip_case_mask = _mm_set1_epi8(flip_case_mask);
for (; src < src_end_sse; src += bytes_sse, dst += bytes_sse)
if (src < src_end_sse)
{
/// load 16 sequential 8-bit characters
const auto chars = _mm_loadu_si128(reinterpret_cast<const __m128i *>(src));
const auto v_not_case_lower_bound = _mm_set1_epi8(not_case_lower_bound - 1);
const auto v_not_case_upper_bound = _mm_set1_epi8(not_case_upper_bound + 1);
const auto v_flip_case_mask = _mm_set1_epi8(flip_case_mask);
/// find which 8-bit sequences belong to range [case_lower_bound, case_upper_bound]
const auto is_not_case
= _mm_and_si128(_mm_cmpgt_epi8(chars, v_not_case_lower_bound), _mm_cmplt_epi8(chars, v_not_case_upper_bound));
for (; src < src_end_sse; src += bytes_sse, dst += bytes_sse)
{
/// load 16 sequential 8-bit characters
const auto chars = _mm_loadu_si128(reinterpret_cast<const __m128i *>(src));
/// keep `flip_case_mask` only where necessary, zero out elsewhere
const auto xor_mask = _mm_and_si128(v_flip_case_mask, is_not_case);
/// find which 8-bit sequences belong to range [case_lower_bound, case_upper_bound]
const auto is_not_case
= _mm_and_si128(_mm_cmpgt_epi8(chars, v_not_case_lower_bound), _mm_cmplt_epi8(chars, v_not_case_upper_bound));
/// flip case by applying calculated mask
const auto cased_chars = _mm_xor_si128(chars, xor_mask);
/// keep `flip_case_mask` only where necessary, zero out elsewhere
const auto xor_mask = _mm_and_si128(v_flip_case_mask, is_not_case);
/// store result back to destination
_mm_storeu_si128(reinterpret_cast<__m128i *>(dst), cased_chars);
/// flip case by applying calculated mask
const auto cased_chars = _mm_xor_si128(chars, xor_mask);
/// store result back to destination
_mm_storeu_si128(reinterpret_cast<__m128i *>(dst), cased_chars);
}
}
#endif