From: Loren Merritt <pengvado@videolan.org>
Date: Sun, 27 Jan 2008 09:12:39 +0000 (+0000)
Subject: faster and smaller x264_macroblock_cache_mv etc
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=75c77579a5e6386250aa85335408d5d8ed475df4;p=libx264

faster and smaller x264_macroblock_cache_mv etc


git-svn-id: svn://svn.videolan.org/x264/trunk@723 df754926-b1dd-0310-bc7b-ec298dee348c
---

diff --git a/common/macroblock.h b/common/macroblock.h
index 570bb80d..29574a80 100644
--- a/common/macroblock.h
+++ b/common/macroblock.h
@@ -294,53 +294,55 @@ void x264_mb_mc( x264_t *h );
 void x264_mb_mc_8x8( x264_t *h, int i8 );
 
 
-static inline void x264_macroblock_cache_ref( x264_t *h, int x, int y, int width, int height, int i_list, int ref )
+static ALWAYS_INLINE void x264_macroblock_cache_rect1( void *dst, int width, int height, uint8_t val )
 {
-    int dy, dx;
-    for( dy = 0; dy < height; dy++ )
+    int dy;
+    if( width == 4 )
     {
-        for( dx = 0; dx < width; dx++ )
-        {
-            h->mb.cache.ref[i_list][X264_SCAN8_0+x+dx+8*(y+dy)] = ref;
-        }
+        uint32_t val2 = val * 0x01010101;
+        for( dy = 0; dy < height; dy++ )
+            ((uint32_t*)dst)[2*dy] = val2;
     }
-}
-static inline void x264_macroblock_cache_mv( x264_t *h, int x, int y, int width, int height, int i_list, int mvx, int mvy )
-{
-    int dy, dx;
-    for( dy = 0; dy < height; dy++ )
+    else // 2
     {
-        for( dx = 0; dx < width; dx++ )
-        {
-            h->mb.cache.mv[i_list][X264_SCAN8_0+x+dx+8*(y+dy)][0] = mvx;
-            h->mb.cache.mv[i_list][X264_SCAN8_0+x+dx+8*(y+dy)][1] = mvy;
-        }
+        uint32_t val2 = val * 0x0101;
+        for( dy = 0; dy < height; dy++ )
+            ((uint16_t*)dst)[4*dy] = val2;
     }
 }
-static inline void x264_macroblock_cache_mvd( x264_t *h, int x, int y, int width, int height, int i_list, int mdx, int mdy )
+static ALWAYS_INLINE void x264_macroblock_cache_rect4( void *dst, int width, int height, uint32_t val )
 {
     int dy, dx;
     for( dy = 0; dy < height; dy++ )
-    {
         for( dx = 0; dx < width; dx++ )
-        {
-            h->mb.cache.mvd[i_list][X264_SCAN8_0+x+dx+8*(y+dy)][0] = mdx;
-            h->mb.cache.mvd[i_list][X264_SCAN8_0+x+dx+8*(y+dy)][1] = mdy;
-        }
-    }
+            ((uint32_t*)dst)[dx+8*dy] = val;
 }
-static inline void x264_macroblock_cache_skip( x264_t *h, int x, int y, int width, int height, int b_skip )
+static ALWAYS_INLINE uint32_t pack16to32( int a, int b )
 {
-    int dy, dx;
-    for( dy = 0; dy < height; dy++ )
-    {
-        for( dx = 0; dx < width; dx++ )
-        {
-            h->mb.cache.skip[X264_SCAN8_0+x+dx+8*(y+dy)] = b_skip;
-        }
-    }
+#ifdef WORDS_BIGENDIAN
+   return (b&0xFFFF) + (a<<16);
+#else
+   return (a&0xFFFF) + (b<<16);
+#endif
+}
+
+static ALWAYS_INLINE void x264_macroblock_cache_ref( x264_t *h, int x, int y, int width, int height, int i_list, uint8_t ref )
+{
+    x264_macroblock_cache_rect1( &h->mb.cache.ref[i_list][X264_SCAN8_0+x+8*y], width, height, ref );
+}
+static ALWAYS_INLINE void x264_macroblock_cache_mv( x264_t *h, int x, int y, int width, int height, int i_list, int mvx, int mvy )
+{
+    x264_macroblock_cache_rect4( &h->mb.cache.mv[i_list][X264_SCAN8_0+x+8*y], width, height, pack16to32(mvx,mvy) );
+}
+static ALWAYS_INLINE void x264_macroblock_cache_mvd( x264_t *h, int x, int y, int width, int height, int i_list, int mdx, int mdy )
+{
+    x264_macroblock_cache_rect4( &h->mb.cache.mvd[i_list][X264_SCAN8_0+x+8*y], width, height, pack16to32(mdx,mdy) );
+}
+static ALWAYS_INLINE void x264_macroblock_cache_skip( x264_t *h, int x, int y, int width, int height, int b_skip )
+{
+    x264_macroblock_cache_rect1( &h->mb.cache.skip[X264_SCAN8_0+x+8*y], width, height, b_skip );
 }
-static inline void x264_macroblock_cache_intra8x8_pred( x264_t *h, int x, int y, int i_mode )
+static ALWAYS_INLINE void x264_macroblock_cache_intra8x8_pred( x264_t *h, int x, int y, int i_mode )
 {
     int *cache = &h->mb.cache.intra4x4_pred_mode[X264_SCAN8_0+x+8*y];
     cache[0] = cache[1] = cache[8] = cache[9] = i_mode;
diff --git a/common/osdep.h b/common/osdep.h
index 9635dd80..eac513ae 100644
--- a/common/osdep.h
+++ b/common/osdep.h
@@ -65,8 +65,10 @@
 
 #if defined(__GNUC__) && (__GNUC__ > 3 || __GNUC__ == 3 && __GNUC_MINOR__ > 0)
 #define UNUSED __attribute__((unused))
+#define ALWAYS_INLINE __attribute__((always_inline)) inline
 #else
 #define UNUSED
+#define ALWAYS_INLINE inline
 #endif
 
 /* threads */