Skip to content
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

Jump to bottom

Avoid masks when possible in AVX2 logic #104

Merged
merged 1 commit into from
Nov 21, 2023
Merged

Avoid masks when possible in AVX2 logic #104

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
14 changes: 4 additions & 10 deletions src/avx2-emu-funcs.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -224,21 +224,18 @@ int avx2_double_compressstore32(void *left_addr,
typename avx2_vector<T>::reg_t reg)
{
using vtype = avx2_vector<T>;
const __m256i oxff = _mm256_set1_epi32(0xFFFFFFFF);

T *leftStore = (T *)left_addr;
T *rightStore = (T *)right_addr;

int32_t shortMask = convert_avx2_mask_to_int(k);
const __m256i &perm = _mm256_loadu_si256(
(const __m256i *)avx2_compressstore_lut32_perm[shortMask].data());
const __m256i &left = _mm256_loadu_si256(
(const __m256i *)avx2_compressstore_lut32_left[shortMask].data());

typename vtype::reg_t temp = vtype::permutevar(reg, perm);

vtype::mask_storeu(leftStore, left, temp);
vtype::mask_storeu(rightStore, _mm256_xor_si256(oxff, left), temp);
vtype::storeu(leftStore, temp);
vtype::storeu(rightStore, temp);

return _mm_popcnt_u32(shortMask);
}
Expand All @@ -250,22 +247,19 @@ int32_t avx2_double_compressstore64(void *left_addr,
typename avx2_vector<T>::reg_t reg)
{
using vtype = avx2_vector<T>;
const __m256i oxff = _mm256_set1_epi32(0xFFFFFFFF);

T *leftStore = (T *)left_addr;
T *rightStore = (T *)right_addr;

int32_t shortMask = convert_avx2_mask_to_int_64bit(k);
const __m256i &perm = _mm256_loadu_si256(
(const __m256i *)avx2_compressstore_lut64_perm[shortMask].data());
const __m256i &left = _mm256_loadu_si256(
(const __m256i *)avx2_compressstore_lut64_left[shortMask].data());

typename vtype::reg_t temp = vtype::cast_from(
_mm256_permutevar8x32_epi32(vtype::cast_to(reg), perm));

vtype::mask_storeu(leftStore, left, temp);
vtype::mask_storeu(rightStore, _mm256_xor_si256(oxff, left), temp);
vtype::storeu(leftStore, temp);
vtype::storeu(rightStore, temp);

return _mm_popcnt_u32(shortMask);
}
Expand Down
Loading