dany/libwebp
1
0
Fork 0
mirror of https://github.com/webmproject/libwebp.git synced 2025-07-15 21:39:59 +02:00
Code Issues Actions Packages Projects Releases Wiki Activity

dsp/enc*: use WEBP_RESTRICT qualifier

Browse Source 
This allows for better vectorization of the C code, inlining of
TrueMotion_SSE2, better load usage in aarch64 and other minor
reordering with ndk r27/gcc-13/clang-16.

This only affects non-vector pointers; any vector pointers are left as a
follow up.

Change-Id: I07e9944d5c0aa5a079b22883ac5a2d649695e4a0
This commit is contained in:
James Zern
2021-07-03 17:59:44 -07:00
parent 201894ef24
commit b1cb37e659
9 changed files with 457 additions and 297 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
src/dsp/enc_sse41.c
View File

@ -23,9 +23,10 @@
//------------------------------------------------------------------------------
// Compute susceptibility based on DCT-coeff histograms.
static void CollectHistogram_SSE41(const uint8_t* ref, const uint8_t* pred,
static void CollectHistogram_SSE41(const uint8_t* WEBP_RESTRICT ref,
const uint8_t* WEBP_RESTRICT pred,
int start_block, int end_block,
VP8Histogram* const histo) {
VP8Histogram* WEBP_RESTRICT const histo) {
const __m128i max_coeff_thresh = _mm_set1_epi16(MAX_COEFF_THRESH);
int j;
int distribution[MAX_COEFF_THRESH + 1] = { 0 };
@ -168,14 +169,16 @@ static int TTransform_SSE41(const uint8_t* inA, const uint8_t* inB,
return sum[0] + sum[1] + sum[2] + sum[3];
}
static int Disto4x4_SSE41(const uint8_t* const a, const uint8_t* const b,
const uint16_t* const w) {
static int Disto4x4_SSE41(const uint8_t* WEBP_RESTRICT const a,
const uint8_t* WEBP_RESTRICT const b,
const uint16_t* WEBP_RESTRICT const w) {
const int diff_sum = TTransform_SSE41(a, b, w);
return abs(diff_sum) >> 5;
}
static int Disto16x16_SSE41(const uint8_t* const a, const uint8_t* const b,
const uint16_t* const w) {
static int Disto16x16_SSE41(const uint8_t* WEBP_RESTRICT const a,
const uint8_t* WEBP_RESTRICT const b,
const uint16_t* WEBP_RESTRICT const w) {
int D = 0;
int x, y;
for (y = 0; y < 16 * BPS; y += 4 * BPS) {
@ -301,17 +304,17 @@ static WEBP_INLINE int DoQuantizeBlock_SSE41(int16_t in[16], int16_t out[16],
#undef PSHUFB_CST
static int QuantizeBlock_SSE41(int16_t in[16], int16_t out[16],
const VP8Matrix* const mtx) {
const VP8Matrix* WEBP_RESTRICT const mtx) {
return DoQuantizeBlock_SSE41(in, out, &mtx->sharpen_[0], mtx);
}
static int QuantizeBlockWHT_SSE41(int16_t in[16], int16_t out[16],
const VP8Matrix* const mtx) {
const VP8Matrix* WEBP_RESTRICT const mtx) {
return DoQuantizeBlock_SSE41(in, out, NULL, mtx);
}
static int Quantize2Blocks_SSE41(int16_t in[32], int16_t out[32],
const VP8Matrix* const mtx) {
const VP8Matrix* WEBP_RESTRICT const mtx) {
int nz;
const uint16_t* const sharpen = &mtx->sharpen_[0];
nz = DoQuantizeBlock_SSE41(in + 0 * 16, out + 0 * 16, sharpen, mtx) << 0;