~2x faster SSE2 RGB24toY, BGR24toY, ARGBToY|UV

global effect is ~2% faster encoding from JPG source
and ~8% faster lossless-webp source decoding to PGM (e.g.)

Also revamped the YUVA case to first accumulate R/G/B value into 16b
temporary buffer, and then doing the UV conversion.
-> New function: WebPConvertRGBA32ToUV

Change-Id: I1d7d0c4003aa02966ad33490ce0fcdc7925cf9f5
This commit is contained in:
Pascal Massimino
2015-11-04 08:51:24 +01:00
committed by James Zern
parent 96f5b4237d
commit bfd3fc02df
4 changed files with 409 additions and 69 deletions

View File

@ -335,10 +335,20 @@ extern void (*WebPConvertARGBToY)(const uint32_t* argb, uint8_t* y, int width);
extern void (*WebPConvertARGBToUV)(const uint32_t* argb, uint8_t* u, uint8_t* v,
int src_width, int do_store);
// Convert a row of accumulated (four-values) of rgba32 toward U/V
extern void (*WebPConvertRGBA32ToUV)(const uint16_t* rgb,
uint8_t* u, uint8_t* v, int width);
// Convert RGB or BGR to Y
extern void (*WebPConvertRGB24ToY)(const uint8_t* rgb, uint8_t* y, int width);
extern void (*WebPConvertBGR24ToY)(const uint8_t* bgr, uint8_t* y, int width);
// used for plain-C fallback.
extern void WebPConvertARGBToUV_C(const uint32_t* argb, uint8_t* u, uint8_t* v,
int src_width, int do_store);
extern void WebPConvertRGBA32ToUV_C(const uint16_t* rgb,
uint8_t* u, uint8_t* v, int width);
// Must be called before using the above.
void WebPInitConvertARGBToYUV(void);