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

Trying simplier white space.

This commit is contained in:
Daniel Lemire 2018-05-30 23:05:18 -04:00
parent f5a0346192
commit 74d1680e7e
1 changed files with 22 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
scalarvssimd/include/avxminifier.h
View File

@ -163,6 +163,7 @@ static inline size_t copy_without_useless_spaces(const uint8_t *buf, size_t len,
_mm_set_epi64x(0ULL, quote_bits), _mm_set1_epi8(0xFF), 0)); _mm_set_epi64x(0ULL, quote_bits), _mm_set1_epi8(0xFF), 0));
quote_mask ^= prev_iter_inside_quote; quote_mask ^= prev_iter_inside_quote;
prev_iter_inside_quote = (uint64_t)((s64)quote_mask >> 63); prev_iter_inside_quote = (uint64_t)((s64)quote_mask >> 63);
/*
const __m256i low_nibble_mask = _mm256_setr_epi8( const __m256i low_nibble_mask = _mm256_setr_epi8(
// 0 9 a b c d // 0 9 a b c d
16, 0, 0, 0, 0, 0, 0, 0, 0, 8, 12, 1, 2, 9, 0, 0, 16, 0, 0, 0, 0, 0, 0, 16, 0, 0, 0, 0, 0, 0, 0, 0, 8, 12, 1, 2, 9, 0, 0, 16, 0, 0, 0, 0, 0, 0,
@ -191,6 +192,27 @@ static inline size_t copy_without_useless_spaces(const uint8_t *buf, size_t len,
uint64_t ws_res_0 = (uint32_t)_mm256_movemask_epi8(tmp_ws_lo); uint64_t ws_res_0 = (uint32_t)_mm256_movemask_epi8(tmp_ws_lo);
uint64_t ws_res_1 = _mm256_movemask_epi8(tmp_ws_hi); uint64_t ws_res_1 = _mm256_movemask_epi8(tmp_ws_hi);
uint64_t whitespace = ~(ws_res_0 | (ws_res_1 << 32)); uint64_t whitespace = ~(ws_res_0 | (ws_res_1 << 32));
*/
__m256i mask_20 = _mm256_set1_epi8(0x20); // c==32
__m256i mask_70 =
_mm256_set1_epi8(0x70); // adding 0x70 does not check low 4-bits
// but moves any value >= 16 above 128
__m256i lut_cntrl = _mm256_setr_epi8(
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xFF, 0xFF, 0x00,
0x00, 0xFF, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0xFF, 0xFF, 0x00, 0x00, 0xFF, 0x00, 0x00);
__m256i tmp_ws_lo = _mm256_or_si256(
_mm256_cmpeq_epi8(mask_20, input_lo),
_mm256_shuffle_epi8(lut_cntrl, _mm256_adds_epu8(mask_70, input_lo)));
__m256i tmp_ws_hi = _mm256_or_si256(
_mm256_cmpeq_epi8(mask_20, input_hi),
_mm256_shuffle_epi8(lut_cntrl, _mm256_adds_epu8(mask_70, input_hi)));
uint64_t ws_res_0 = (uint32_t)_mm256_movemask_epi8(tmp_ws_lo);
uint64_t ws_res_1 = _mm256_movemask_epi8(tmp_ws_hi);
uint64_t whitespace = (ws_res_0 | (ws_res_1 << 32));
whitespace &= ~quote_mask; whitespace &= ~quote_mask;
// //