jasder
/
simdjson
1
0
Fork 0
Code Issues Pull Requests Releases 2 Wiki Activity

Trimming the table approach.

This commit is contained in:
Daniel Lemire 2018-07-20 09:20:34 -04:00
parent 41d2914a3b
commit a65066e962
2 changed files with 9 additions and 33036 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
scalarvssimd/include/avxminifier.h
View File

@ -117,11 +117,11 @@ static inline size_t copy_without_useless_spaces_avx(const uint8_t *buf, size_t
int pop3 = _popcnt64((~whitespace) & UINT64_C(0xFFFFFFFFFFFF)); int pop3 = _popcnt64((~whitespace) & UINT64_C(0xFFFFFFFFFFFF));
int pop4 = _popcnt64((~whitespace)); int pop4 = _popcnt64((~whitespace));
__m256i vmask1 = __m256i vmask1 =
_mm256_loadu2_m128i((const __m128i *)mask128_epi8 + mask2, _mm256_loadu2_m128i((const __m128i *)mask128_epi8 + (mask2 & 0x7FFFF),
(const __m128i *)mask128_epi8 + mask1); (const __m128i *)mask128_epi8 + (mask1 & 0x7FFFF));
__m256i vmask2 = __m256i vmask2 =
_mm256_loadu2_m128i((const __m128i *)mask128_epi8 + mask4, _mm256_loadu2_m128i((const __m128i *)mask128_epi8 + (mask4 & 0x7FFFF),
(const __m128i *)mask128_epi8 + mask3); (const __m128i *)mask128_epi8 + (mask3 & 0x7FFFF));
__m256i result1 = _mm256_shuffle_epi8(input_lo, vmask1); __m256i result1 = _mm256_shuffle_epi8(input_lo, vmask1);
__m256i result2 = _mm256_shuffle_epi8(input_hi, vmask2); __m256i result2 = _mm256_shuffle_epi8(input_hi, vmask2);
_mm256_storeu2_m128i((__m128i *)(out + pop1), (__m128i *)out, result1); _mm256_storeu2_m128i((__m128i *)(out + pop1), (__m128i *)out, result1);
@ -199,10 +199,10 @@ static inline size_t copy_without_useless_spaces_avx(const uint8_t *buf, size_t
int pop2 = _popcnt64((~whitespace) & UINT64_C(0xFFFFFFFF)); int pop2 = _popcnt64((~whitespace) & UINT64_C(0xFFFFFFFF));
int pop3 = _popcnt64((~whitespace) & UINT64_C(0xFFFFFFFFFFFF)); int pop3 = _popcnt64((~whitespace) & UINT64_C(0xFFFFFFFFFFFF));
int pop4 = _popcnt64((~whitespace)); int pop4 = _popcnt64((~whitespace));
__m256i vmask1 = _mm256_loadu2_m128i((const __m128i *)mask128_epi8 + mask2, __m256i vmask1 = _mm256_loadu2_m128i((const __m128i *)mask128_epi8 + (mask2 & 0x7FFF),
(const __m128i *)mask128_epi8 + mask1); (const __m128i *)mask128_epi8 + (mask1 & 0x7FFF));
__m256i vmask2 = _mm256_loadu2_m128i((const __m128i *)mask128_epi8 + mask4, __m256i vmask2 = _mm256_loadu2_m128i((const __m128i *)mask128_epi8 + (mask4 & 0x7FFF),
(const __m128i *)mask128_epi8 + mask3); (const __m128i *)mask128_epi8 + (mask3 & 0x7FFF));
__m256i result1 = _mm256_shuffle_epi8(input_lo, vmask1); __m256i result1 = _mm256_shuffle_epi8(input_lo, vmask1);
__m256i result2 = _mm256_shuffle_epi8(input_hi, vmask2); __m256i result2 = _mm256_shuffle_epi8(input_hi, vmask2);
_mm256_storeu2_m128i((__m128i *)(buffer + pop1), (__m128i *)buffer, _mm256_storeu2_m128i((__m128i *)(buffer + pop1), (__m128i *)buffer,

33029
scalarvssimd/include/simdprune_tables.h
View File

File diff suppressed because it is too large Load Diff