// Copyright 2014 Google Inc. All Rights Reserved. // // Use of this source code is governed by a BSD-style license // that can be found in the COPYING file in the root of the source // tree. An additional intellectual property rights grant can be found // in the file PATENTS. All contributing project authors may // be found in the AUTHORS file in the root of the source tree. // ----------------------------------------------------------------------------- // // SSE2 variant of methods for lossless decoder // // Author: Skal ([email protected]) #include "src/dsp/dsp.h" #if defined(WEBP_USE_SSE2) #include "src/dsp/common_sse2.h" #include "src/dsp/lossless.h" #include "src/dsp/lossless_common.h" #include <emmintrin.h> //------------------------------------------------------------------------------ // Predictor Transform static WEBP_INLINE uint32_t ClampedAddSubtractFull_SSE2(uint32_t c0, uint32_t c1, uint32_t c2) { … } static WEBP_INLINE uint32_t ClampedAddSubtractHalf_SSE2(uint32_t c0, uint32_t c1, uint32_t c2) { … } static WEBP_INLINE uint32_t Select_SSE2(uint32_t a, uint32_t b, uint32_t c) { … } static WEBP_INLINE void Average2_m128i(const __m128i* const a0, const __m128i* const a1, __m128i* const avg) { … } static WEBP_INLINE void Average2_uint32_SSE2(const uint32_t a0, const uint32_t a1, __m128i* const avg) { … } static WEBP_INLINE __m128i Average2_uint32_16_SSE2(uint32_t a0, uint32_t a1) { … } static WEBP_INLINE uint32_t Average2_SSE2(uint32_t a0, uint32_t a1) { … } static WEBP_INLINE uint32_t Average3_SSE2(uint32_t a0, uint32_t a1, uint32_t a2) { … } static WEBP_INLINE uint32_t Average4_SSE2(uint32_t a0, uint32_t a1, uint32_t a2, uint32_t a3) { … } static uint32_t Predictor5_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor6_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor7_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor8_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor9_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor10_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor11_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor12_SSE2(const uint32_t* const left, const uint32_t* const top) { … } static uint32_t Predictor13_SSE2(const uint32_t* const left, const uint32_t* const top) { … } // Batch versions of those functions. // Predictor0: ARGB_BLACK. static void PredictorAdd0_SSE2(const uint32_t* in, const uint32_t* upper, int num_pixels, uint32_t* out) { … } // Predictor1: left. static void PredictorAdd1_SSE2(const uint32_t* in, const uint32_t* upper, int num_pixels, uint32_t* out) { … } // Macro that adds 32-bit integers from IN using mod 256 arithmetic // per 8 bit channel. #define GENERATE_PREDICTOR_1 … // Predictor2: Top. GENERATE_PREDICTOR_1 // Predictor3: Top-right. GENERATE_PREDICTOR_1 // Predictor4: Top-left. GENERATE_PREDICTOR_1 #undef GENERATE_PREDICTOR_1 // Due to averages with integers, values cannot be accumulated in parallel for // predictors 5 to 7. GENERATE_PREDICTOR_ADD(Predictor5_SSE2, PredictorAdd5_SSE2) GENERATE_PREDICTOR_ADD(Predictor6_SSE2, PredictorAdd6_SSE2) GENERATE_PREDICTOR_ADD(Predictor7_SSE2, PredictorAdd7_SSE2) #define GENERATE_PREDICTOR_2 … // Predictor8: average TL T. GENERATE_PREDICTOR_2 // Predictor9: average T TR. GENERATE_PREDICTOR_2 #undef GENERATE_PREDICTOR_2 // Predictor10: average of (average of (L,TL), average of (T, TR)). #define DO_PRED10 … #define DO_PRED10_SHIFT … static void PredictorAdd10_SSE2(const uint32_t* in, const uint32_t* upper, int num_pixels, uint32_t* out) { … } #undef DO_PRED10 #undef DO_PRED10_SHIFT // Predictor11: select. #define DO_PRED11 … #define DO_PRED11_SHIFT … static void PredictorAdd11_SSE2(const uint32_t* in, const uint32_t* upper, int num_pixels, uint32_t* out) { … } #undef DO_PRED11 #undef DO_PRED11_SHIFT // Predictor12: ClampedAddSubtractFull. #define DO_PRED12 … #define DO_PRED12_SHIFT … static void PredictorAdd12_SSE2(const uint32_t* in, const uint32_t* upper, int num_pixels, uint32_t* out) { … } #undef DO_PRED12 #undef DO_PRED12_SHIFT // Due to averages with integers, values cannot be accumulated in parallel for // predictors 13. GENERATE_PREDICTOR_ADD(Predictor13_SSE2, PredictorAdd13_SSE2) //------------------------------------------------------------------------------ // Subtract-Green Transform static void AddGreenToBlueAndRed_SSE2(const uint32_t* const src, int num_pixels, uint32_t* dst) { … } //------------------------------------------------------------------------------ // Color Transform static void TransformColorInverse_SSE2(const VP8LMultipliers* const m, const uint32_t* const src, int num_pixels, uint32_t* dst) { … } //------------------------------------------------------------------------------ // Color-space conversion functions static void ConvertBGRAToRGB_SSE2(const uint32_t* src, int num_pixels, uint8_t* dst) { … } static void ConvertBGRAToRGBA_SSE2(const uint32_t* src, int num_pixels, uint8_t* dst) { … } static void ConvertBGRAToRGBA4444_SSE2(const uint32_t* src, int num_pixels, uint8_t* dst) { … } static void ConvertBGRAToRGB565_SSE2(const uint32_t* src, int num_pixels, uint8_t* dst) { … } static void ConvertBGRAToBGR_SSE2(const uint32_t* src, int num_pixels, uint8_t* dst) { … } //------------------------------------------------------------------------------ // Entry point extern void VP8LDspInitSSE2(void); WEBP_TSAN_IGNORE_FUNCTION void VP8LDspInitSSE2(void) { … } #else // !WEBP_USE_SSE2 WEBP_DSP_INIT_STUB(VP8LDspInitSSE2) #endif // WEBP_USE_SSE2
Codebase Browser
godot