hashirama/uvg266
1
0
Fork 0
mirror of https://github.com/ultravideo/uvg266.git synced 2024-11-27 19:24:06 +00:00
Code Issues Actions Projects Releases Wiki Activity

Tidy pack_16x16b_to_16x2b comments

Browse source
This commit is contained in:
Pauli Oikkonen 2019-01-03 16:37:05 +02:00
parent 016eb014ad
commit 45ac6e6d03
1 changed files with 5 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
src/strategies/avx2/encode_coding_tree-avx2.c
View file

@ -52,15 +52,14 @@ static INLINE uint32_t pack_16x16b_to_16x2b(__m256i src)
{
/*
* For each 16-bit element in src:
* Clip it to max. 3 (assuming the numbers to be unsigned)
* Shift
* 0000 0000 0000 00XY Clip to [0, 3]
* ABCD EFGH IJKL MNOP Original elements
* 0000 0000 0000 00XY Element clipped to [0, 3] using _mm256_min_epu16
* 0000 000X Y000 0000 Shift word to align LSBs across byte boundary
* 0000 0001 1000 0000 cmpmask to be compared against
* 0000 0001 1000 0000 Comparison mask to be compared against
* XXXX XXXX YYYY YYYY Comparison result, for movemask
*/
const __m256i threes = _mm256_set1_epi16(3);
const __m256i cmpmask = _mm256_set1_epi16(0x0180);
const __m256i threes = _mm256_set1_epi16 (3);
const __m256i cmpmask = _mm256_set1_epi16 (0x0180);
__m256i clipped = _mm256_min_epu16 (src, threes);
__m256i shifted = _mm256_slli_epi16 (clipped, 7);