Skip to content

Improve vectorized string::find_meow_of for small bitmap cases #6080

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
AlexGuteniev wants to merge 4 commits into
base: main
Choose a base branch
from
+16 −9
Open

Improve vectorized string::find_meow_of for small bitmap cases #6080

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b6084c7
Better small bitmap algorithm
AlexGuteniev Feb 8, 2026
aec5158
Clarity
AlexGuteniev Feb 8, 2026
0c7fa62
Format
AlexGuteniev Feb 8, 2026
ba15737
Improve comments.
StephanTLavavej Feb 8, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 12 additions & 9 deletions stl/src/vector_algorithms.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5645,25 +5645,28 @@ namespace {
}

template <class _Ty>
__m256i _Make_bitmap_small(const _Ty* _Needle_ptr, const size_t _Needle_length) noexcept {
__forceinline __m256i _Make_bitmap_small(const _Ty* _Needle_ptr, const size_t _Needle_length) noexcept {
__m256i _Bitmap = _mm256_setzero_si256();

const _Ty* const _Stop = _Needle_ptr + _Needle_length;

const __m256i _Pow = _mm256_set_epi32(7 << 5, 6 << 5, 5 << 5, 4 << 5, 3 << 5, 2 << 5, 1 << 5, 0 << 5);
const __m256i _Ones = _mm256_set1_epi32(1);

for (; _Needle_ptr != _Stop; ++_Needle_ptr) {
const _Ty _Val = *_Needle_ptr;
const __m128i _Count_low = _mm_cvtsi32_si128(_Val & 0x3F);
const auto _Count_high_x8 = static_cast<uint32_t>((_Val >> 3) & 0x18);
const __m256i _One_1_high = _mm256_cvtepu8_epi64(_mm_cvtsi32_si128(1u << _Count_high_x8));
const __m256i _One_1 = _mm256_sll_epi64(_One_1_high, _Count_low);
_Bitmap = _mm256_or_si256(_Bitmap, _One_1);
const _Ty _Val = *_Needle_ptr;
const __m128i _Count_low = _mm_cvtsi32_si128(static_cast<uint32_t>(_Val));
const __m256i _Count_all = _mm256_broadcastd_epi32(_Count_low);
const __m256i _Count_one = _mm256_xor_si256(_Pow, _Count_all);
const __m256i _One_1 = _mm256_sllv_epi32(_Ones, _Count_one);
_Bitmap = _mm256_or_si256(_Bitmap, _One_1);
}

return _Bitmap;
}

template <class _Ty>
__m256i _Make_bitmap_large(const _Ty* _Needle_ptr, const size_t _Needle_length) noexcept {
__forceinline __m256i _Make_bitmap_large(const _Ty* _Needle_ptr, const size_t _Needle_length) noexcept {
alignas(32) uint8_t _Table[256] = {};

const _Ty* const _Stop = _Needle_ptr + _Needle_length;
Expand All @@ -5686,7 +5689,7 @@ namespace {
}

template <class _Ty>
__m256i _Make_bitmap(const _Ty* const _Needle_ptr, const size_t _Needle_length) noexcept {
__forceinline __m256i _Make_bitmap(const _Ty* const _Needle_ptr, const size_t _Needle_length) noexcept {
if (_Needle_length <= 20) {
return _Make_bitmap_small(_Needle_ptr, _Needle_length);
} else {
Expand Down