python-pillow/Pillow
1
1
Fork 0
mirror of https://github.com/python-pillow/Pillow.git synced 2025-08-20 20:24:45 +03:00
Code Issues Packages Projects Releases Wiki Activity

SIMD ColorLUT. remove SHIFT_ROUNDING

Browse Source
This commit is contained in:
Alexander 2018-03-27 03:51:21 +03:00 committed by Alexander Karpinsky
parent 16f69404f1
commit f50dba4aad
1 changed files with 18 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
src/libImaging/ColorLUT.c
View File

@ -13,8 +13,7 @@
#define SCALE_BITS (32 - 8 - 6) #define SCALE_BITS (32 - 8 - 6)
#define SCALE_MASK ((1 << SCALE_BITS) - 1) #define SCALE_MASK ((1 << SCALE_BITS) - 1)
#define SHIFT_BITS (16 - 2) #define SHIFT_BITS (16 - 1)
#define SHIFT_ROUNDING (1<<(SHIFT_BITS-1))
static inline UINT8 static inline UINT8
clip8(int in) { clip8(int in) {
@ -82,7 +81,6 @@ ImagingColorLUT3D_linear(
(size1D - 1) / 255.0 * (1<<SCALE_BITS)); (size1D - 1) / 255.0 * (1<<SCALE_BITS));
__m128i scale_mask = _mm_set1_epi32(SCALE_MASK); __m128i scale_mask = _mm_set1_epi32(SCALE_MASK);
__m128i index_mul = _mm_set_epi32(0, size1D_2D, size1D, 1); __m128i index_mul = _mm_set_epi32(0, size1D_2D, size1D, 1);
__m128i shift_rounding = _mm_set1_epi32(SHIFT_ROUNDING);
__m128i shuffle_source = _mm_set_epi8(-1,-1, -1,-1, 11,10, 5,4, 9,8, 3,2, 7,6, 1,0); __m128i shuffle_source = _mm_set_epi8(-1,-1, -1,-1, 11,10, 5,4, 9,8, 3,2, 7,6, 1,0);
__m128i left_mask = _mm_set1_epi32(0x0000ffff); __m128i left_mask = _mm_set1_epi32(0x0000ffff);
__m128i right_mask = _mm_set1_epi32(0xffff0000); __m128i right_mask = _mm_set1_epi32(0xffff0000);
@ -122,7 +120,7 @@ ImagingColorLUT3D_linear(
__m128i leftleft, leftright, rightleft, rightright; __m128i leftleft, leftright, rightleft, rightright;
shift = _mm_or_si128( shift = _mm_or_si128(
_mm_sub_epi32(_mm_set1_epi32(1<<SHIFT_BITS), shift), _mm_sub_epi32(_mm_set1_epi32((1<<SHIFT_BITS)-1), shift),
_mm_slli_epi32(shift, 16)); _mm_slli_epi32(shift, 16));
shift1D = _mm_shuffle_epi32(shift, 0x00); shift1D = _mm_shuffle_epi32(shift, 0x00);
@ -132,39 +130,37 @@ ImagingColorLUT3D_linear(
if (table_channels == 3) { if (table_channels == 3) {
source = _mm_shuffle_epi8( source = _mm_shuffle_epi8(
_mm_loadu_si128((__m128i *) &table[idx + 0]), shuffle_source); _mm_loadu_si128((__m128i *) &table[idx + 0]), shuffle_source);
leftleft = _mm_and_si128(_mm_srai_epi32(_mm_add_epi32(_mm_madd_epi16( leftleft = _mm_and_si128(_mm_srai_epi32(_mm_madd_epi16(
source, shift1D), shift_rounding), SHIFT_BITS), left_mask); source, shift1D), SHIFT_BITS), left_mask);
source = _mm_shuffle_epi8( source = _mm_shuffle_epi8(
_mm_loadu_si128((__m128i *) &table[idx + size1D*3]), shuffle_source); _mm_loadu_si128((__m128i *) &table[idx + size1D*3]), shuffle_source);
leftright = _mm_and_si128(_mm_slli_epi32(_mm_add_epi32(_mm_madd_epi16( leftright = _mm_and_si128(_mm_slli_epi32(_mm_madd_epi16(
source, shift1D), shift_rounding), 16 - SHIFT_BITS), right_mask); source, shift1D), 16 - SHIFT_BITS), right_mask);
source = _mm_shuffle_epi8( source = _mm_shuffle_epi8(
_mm_loadu_si128((__m128i *) &table[idx + size1D_2D*3]), shuffle_source); _mm_loadu_si128((__m128i *) &table[idx + size1D_2D*3]), shuffle_source);
rightleft = _mm_and_si128(_mm_srai_epi32(_mm_add_epi32(_mm_madd_epi16( rightleft = _mm_and_si128(_mm_srai_epi32(_mm_madd_epi16(
source, shift1D), shift_rounding), SHIFT_BITS), left_mask); source, shift1D), SHIFT_BITS), left_mask);
source = _mm_shuffle_epi8( source = _mm_shuffle_epi8(
_mm_loadu_si128((__m128i *) &table[idx + size1D_2D*3 + size1D*3]), shuffle_source); _mm_loadu_si128((__m128i *) &table[idx + size1D_2D*3 + size1D*3]), shuffle_source);
rightright = _mm_and_si128(_mm_slli_epi32(_mm_add_epi32(_mm_madd_epi16( rightright = _mm_and_si128(_mm_slli_epi32(_mm_madd_epi16(
source, shift1D), shift_rounding), 16 - SHIFT_BITS), right_mask); source, shift1D), 16 - SHIFT_BITS), right_mask);
left = _mm_and_si128(_mm_srai_epi32(_mm_add_epi32(_mm_madd_epi16( left = _mm_and_si128(_mm_srai_epi32(_mm_madd_epi16(
_mm_or_si128(leftleft, leftright), shift2D), _mm_or_si128(leftleft, leftright), shift2D),
shift_rounding), SHIFT_BITS), left_mask); SHIFT_BITS), left_mask);
right = _mm_and_si128(_mm_slli_epi32(_mm_add_epi32(_mm_madd_epi16( right = _mm_and_si128(_mm_slli_epi32(_mm_madd_epi16(
_mm_or_si128(rightleft, rightright), shift2D), _mm_or_si128(rightleft, rightright), shift2D),
shift_rounding), 16 - SHIFT_BITS), right_mask); 16 - SHIFT_BITS), right_mask);
result = _mm_srai_epi32(_mm_add_epi32(_mm_madd_epi16( result = _mm_madd_epi16(_mm_or_si128(left, right), shift3D);
_mm_or_si128(left, right), shift3D),
shift_rounding), SHIFT_BITS);
result = _mm_srai_epi32( result = _mm_srai_epi32(_mm_add_epi32(
_mm_add_epi32(_mm_set1_epi32(PRECISION_ROUNDING), result), _mm_set1_epi32(PRECISION_ROUNDING<<SHIFT_BITS), result),
PRECISION_BITS); PRECISION_BITS + SHIFT_BITS);
result = _mm_packs_epi32(result, result); result = _mm_packs_epi32(result, result);
rowOut[x] = _mm_cvtsi128_si32(_mm_packus_epi16(result, result)); rowOut[x] = _mm_cvtsi128_si32(_mm_packus_epi16(result, result));