Skip to content

Instantly share code, notes, and snippets.

@geofflangdale
Created March 6, 2019 00:45
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save geofflangdale/1a73c9091e04267548bf84baab018596 to your computer and use it in GitHub Desktop.
Save geofflangdale/1a73c9091e04267548bf84baab018596 to your computer and use it in GitHub Desktop.
Finding quote pairs with PCLMULQDQ
really_inline uint64_t find_quote_mask_and_bits(
__m256i input_lo, __m256i input_hi, uint64_t odd_ends,
uint64_t &prev_iter_inside_quote, uint64_t &quote_bits) {
quote_bits =
cmp_mask_against_input(input_lo, input_hi, _mm256_set1_epi8('"'));
quote_bits = quote_bits & ~odd_ends;
uint64_t quote_mask = _mm_cvtsi128_si64(_mm_clmulepi64_si128(
_mm_set_epi64x(0ULL, quote_bits), _mm_set1_epi8(0xFF), 0));
quote_mask ^= prev_iter_inside_quote;
// right shift of a signed value expected to be well-defined and standard
// compliant as of C++20,
// John Regher from Utah U. says this is fine code
prev_iter_inside_quote =
static_cast<uint64_t>(static_cast<int64_t>(quote_mask) >> 63);
return quote_mask;
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment