[FFmpeg-devel] [PATCH v2 1/2] swscale/input: unify grayf32 funcs with rgbf32 funcs

[mindmark at gmail.com]

From: Mark Reid <mindmark at gmail.com>

This is ment to be a cosmetic change

old timings:
  42780 UNITS in grayf32le,       1 runs,      0 skips
  56720 UNITS in grayf32le,       2 runs,      0 skips
  67265 UNITS in grayf32le,       4 runs,      0 skips
  58082 UNITS in grayf32le,       8 runs,      0 skips
  63512 UNITS in grayf32le,      16 runs,      0 skips
  52720 UNITS in grayf32le,      32 runs,      0 skips
  46491 UNITS in grayf32le,      64 runs,      0 skips

  68500 UNITS in grayf32be,       1 runs,      0 skips
  66930 UNITS in grayf32be,       2 runs,      0 skips
  62305 UNITS in grayf32be,       4 runs,      0 skips
  55510 UNITS in grayf32be,       8 runs,      0 skips
  50216 UNITS in grayf32be,      16 runs,      0 skips
  44480 UNITS in grayf32be,      32 runs,      0 skips
  42394 UNITS in grayf32be,      64 runs,      0 skips

new timings:
  46660 UNITS in grayf32le,       1 runs,      0 skips
  51830 UNITS in grayf32le,       2 runs,      0 skips
  53390 UNITS in grayf32le,       4 runs,      0 skips
  50910 UNITS in grayf32le,       8 runs,      0 skips
  44968 UNITS in grayf32le,      16 runs,      0 skips
  40349 UNITS in grayf32le,      32 runs,      0 skips
  38330 UNITS in grayf32le,      64 runs,      0 skips

  39980 UNITS in grayf32be,       1 runs,      0 skips
  49630 UNITS in grayf32be,       2 runs,      0 skips
  53540 UNITS in grayf32be,       4 runs,      0 skips
  59767 UNITS in grayf32be,       8 runs,      0 skips
  51206 UNITS in grayf32be,      16 runs,      0 skips
  44743 UNITS in grayf32be,      32 runs,      0 skips
  41468 UNITS in grayf32be,      64 runs,      0 skips
---
 libswscale/input.c | 36 +++++++++++-------------------------
 1 file changed, 11 insertions(+), 25 deletions(-)

diff --git a/libswscale/input.c b/libswscale/input.c
index 336f957c8c..90efdd2ffc 100644
--- a/libswscale/input.c
+++ b/libswscale/input.c
@@ -1013,31 +1013,19 @@ static av_always_inline void planar_rgbf32_to_y(uint8_t *_dst, const uint8_t *_s
     }
 }

-#undef rdpx
-
 static av_always_inline void grayf32ToY16_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused1,
-                                            const uint8_t *unused2, int width, uint32_t *unused)
+                                            const uint8_t *unused2, int width, int is_be, uint32_t *unused)
 {
     int i;
     const float *src = (const float *)_src;
     uint16_t *dst    = (uint16_t *)_dst;

     for (i = 0; i < width; ++i){
-        dst[i] = av_clip_uint16(lrintf(65535.0f * src[i]));
+        dst[i] = av_clip_uint16(lrintf(65535.0f * rdpx(src + i)));
     }
 }

-static av_always_inline void grayf32ToY16_bswap_c(uint8_t *_dst, const uint8_t *_src, const uint8_t *unused1,
-                                                  const uint8_t *unused2, int width, uint32_t *unused)
-{
-    int i;
-    const uint32_t *src = (const uint32_t *)_src;
-    uint16_t *dst    = (uint16_t *)_dst;
-
-    for (i = 0; i < width; ++i){
-        dst[i] = av_clip_uint16(lrintf(65535.0f * av_int2float(av_bswap32(src[i]))));
-    }
-}
+#undef rdpx

 #define rgb9plus_planar_funcs_endian(nbits, endian_name, endian)                                    \
 static void planar_rgb##nbits##endian_name##_to_y(uint8_t *dst, const uint8_t *src[4],              \
@@ -1092,6 +1080,12 @@ static void planar_rgbf32##endian_name##_to_a(uint8_t *dst, const uint8_t *src[4
                                               int w, int32_t *rgb2yuv)                              \
 {                                                                                                   \
     planar_rgbf32_to_a(dst, src, w, endian, rgb2yuv);                                               \
+}                                                                                                   \
+static void grayf32##endian_name##ToY16_c(uint8_t *dst, const uint8_t *src,                         \
+                                          const uint8_t *unused1, const uint8_t *unused2,           \
+                                          int width, uint32_t *unused)                              \
+{                                                                                                   \
+    grayf32ToY16_c(dst, src, unused1, unused2, width, endian, unused);                              \
 }

 rgbf32_planar_funcs_endian(le, 0)
@@ -1699,18 +1693,10 @@ av_cold void ff_sws_init_input_funcs(SwsContext *c)
         c->lumToYV12 = p010BEToY_c;
         break;
     case AV_PIX_FMT_GRAYF32LE:
-#if HAVE_BIGENDIAN
-        c->lumToYV12 = grayf32ToY16_bswap_c;
-#else
-        c->lumToYV12 = grayf32ToY16_c;
-#endif
+        c->lumToYV12 = grayf32leToY16_c;
         break;
     case AV_PIX_FMT_GRAYF32BE:
-#if HAVE_BIGENDIAN
-        c->lumToYV12 = grayf32ToY16_c;
-#else
-        c->lumToYV12 = grayf32ToY16_bswap_c;
-#endif
+        c->lumToYV12 = grayf32beToY16_c;
         break;
     case AV_PIX_FMT_Y210LE:
         c->lumToYV12 = y210le_Y_c;
--
2.31.1.windows.1