mmx: add src_8888_0565

Uses the pmadd technique described in
http://software.intel.com/sites/landingpage/legacy/mmx/MMX_App_24-16_Bit_Conversion.pdf

The technique uses the packssdw instruction which uses signed
saturatation. This works in their example because they pack 888 to 555
leaving the high bit as zero. For packing to 565, it is unsuitable, so
we replace it with an or+shuffle.

Loongson:
src_8888_0565 =  L1: 106.13  L2:  83.57  M: 33.46 ( 68.90%)  HT: 30.29  VT: 27.67  R: 26.11  RT: 15.06 ( 135Kops/s)
src_8888_0565 =  L1: 122.10  L2: 117.53  M: 37.97 ( 78.58%)  HT: 33.14  VT: 30.09  R: 29.01  RT: 15.76 ( 139Kops/s)

ARM/iwMMXt:
src_8888_0565 =  L1:  67.88  L2:  56.61  M: 31.20 ( 56.74%)  HT: 29.22  VT: 27.01  R: 25.39  RT: 19.29 ( 130Kops/s)
src_8888_0565 =  L1: 110.38  L2:  82.33  M: 40.92 ( 73.22%)  HT: 35.63  VT: 32.22  R: 30.07  RT: 18.40 ( 132Kops/s)

2 files changed, 96 insertions, 0 deletions

diff --git a/pixman/loongson-mmintrin.h b/pixman/loongson-mmintrin.h
index 76ae8920..8295ba01 100644
--- a/pixman/loongson-mmintrin.h
+++ b/pixman/loongson-mmintrin.h
@@ -84,6 +84,17 @@ _mm_empty (void)
 }
 
 extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_madd_pi16 (__m64 __m1, __m64 __m2)
+{
+	__m64 ret;
+	asm("pmaddhw %0, %1, %2\n\t"
+	   : "=f" (ret)
+	   : "f" (__m1), "f" (__m2)
+	);
+	return ret;
+}
+
+extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
 _mm_mulhi_pu16 (__m64 __m1, __m64 __m2)
 {
 	__m64 ret;
diff --git a/pixman/pixman-mmx.c b/pixman/pixman-mmx.c
index 320e20af..7fe19d53 100644
--- a/pixman/pixman-mmx.c
+++ b/pixman/pixman-mmx.c
@@ -179,9 +179,12 @@ typedef struct
     mmxdatafield mmx_4x0080;
     mmxdatafield mmx_565_rgb;
     mmxdatafield mmx_565_unpack_multiplier;
+    mmxdatafield mmx_565_pack_multiplier;
     mmxdatafield mmx_565_r;
     mmxdatafield mmx_565_g;
     mmxdatafield mmx_565_b;
+    mmxdatafield mmx_packed_565_rb;
+    mmxdatafield mmx_packed_565_g;
 #ifndef USE_LOONGSON_MMI
     mmxdatafield mmx_mask_0;
     mmxdatafield mmx_mask_1;
@@ -207,9 +210,12 @@ static const mmx_data_t c =
     MMXDATA_INIT (.mmx_4x0080,                   0x0080008000800080),
     MMXDATA_INIT (.mmx_565_rgb,                  0x000001f0003f001f),
     MMXDATA_INIT (.mmx_565_unpack_multiplier,    0x0000008404100840),
+    MMXDATA_INIT (.mmx_565_pack_multiplier,      0x2000000420000004),
     MMXDATA_INIT (.mmx_565_r,                    0x000000f800000000),
     MMXDATA_INIT (.mmx_565_g,                    0x0000000000fc0000),
     MMXDATA_INIT (.mmx_565_b,                    0x00000000000000f8),
+    MMXDATA_INIT (.mmx_packed_565_rb,            0x00f800f800f800f8),
+    MMXDATA_INIT (.mmx_packed_565_g,             0x0000fc000000fc00),
 #ifndef USE_LOONGSON_MMI
     MMXDATA_INIT (.mmx_mask_0,                   0xffffffffffff0000),
     MMXDATA_INIT (.mmx_mask_1,                   0xffffffff0000ffff),
@@ -567,6 +573,27 @@ pack_565 (__m64 pixel, __m64 target, int pos)
 #endif
 }
 
+static force_inline __m64
+pack_4xpacked565 (__m64 a, __m64 b)
+{
+    __m64 rb0 = _mm_and_si64 (a, MC (packed_565_rb));
+    __m64 rb1 = _mm_and_si64 (b, MC (packed_565_rb));
+
+    __m64 t0 = _mm_madd_pi16 (rb0, MC (565_pack_multiplier));
+    __m64 t1 = _mm_madd_pi16 (rb1, MC (565_pack_multiplier));
+
+    __m64 g0 = _mm_and_si64 (a, MC (packed_565_g));
+    __m64 g1 = _mm_and_si64 (b, MC (packed_565_g));
+
+    t0 = _mm_or_si64 (t0, g0);
+    t1 = _mm_or_si64 (t1, g1);
+
+    t0 = shift(t0, -5);
+    t1 = shift(t1, -5 + 16);
+
+    return _mm_shuffle_pi16 (_mm_or_si64 (t0, t1), _MM_SHUFFLE (3, 1, 2, 0));
+}
+
 #ifndef _MSC_VER
 
 static force_inline __m64
@@ -2091,6 +2118,60 @@ pixman_fill_mmx (uint32_t *bits,
 }
 
 static void
+mmx_composite_src_x888_0565 (pixman_implementation_t *imp,
+                             pixman_composite_info_t *info)
+{
+    PIXMAN_COMPOSITE_ARGS (info);
+    uint16_t    *dst_line, *dst;
+    uint32_t    *src_line, *src, s;
+    int dst_stride, src_stride;
+    int32_t w;
+
+    PIXMAN_IMAGE_GET_LINE (src_image, src_x, src_y, uint32_t, src_stride, src_line, 1);
+    PIXMAN_IMAGE_GET_LINE (dest_image, dest_x, dest_y, uint16_t, dst_stride, dst_line, 1);
+
+    while (height--)
+    {
+	dst = dst_line;
+	dst_line += dst_stride;
+	src = src_line;
+	src_line += src_stride;
+	w = width;
+
+	while (w && (unsigned long)dst & 7)
+	{
+	    s = *src++;
+	    *dst = CONVERT_8888_TO_0565 (s);
+	    dst++;
+	    w--;
+	}
+
+	while (w >= 4)
+	{
+	    __m64 vdest;
+	    __m64 vsrc0 = ldq_u ((__m64 *)(src + 0));
+	    __m64 vsrc1 = ldq_u ((__m64 *)(src + 2));
+
+	    vdest = pack_4xpacked565 (vsrc0, vsrc1);
+
+	    *(__m64 *)dst = vdest;
+
+	    w -= 4;
+	    src += 4;
+	    dst += 4;
+	}
+
+	while (w)
+	{
+	    s = *src++;
+	    *dst = CONVERT_8888_TO_0565 (s);
+	    dst++;
+	    w--;
+	}
+    }
+}
+
+static void
 mmx_composite_src_n_8_8888 (pixman_implementation_t *imp,
                             pixman_composite_info_t *info)
 {
@@ -3433,6 +3514,10 @@ static const pixman_fast_path_t mmx_fast_paths[] =
     PIXMAN_STD_FAST_PATH    (ADD,  a8,       null,     a8,       mmx_composite_add_8_8		   ),
     PIXMAN_STD_FAST_PATH    (ADD,  solid,    a8,       a8,       mmx_composite_add_n_8_8           ),
 
+    PIXMAN_STD_FAST_PATH    (SRC,  a8r8g8b8, null,     r5g6b5,   mmx_composite_src_x888_0565       ),
+    PIXMAN_STD_FAST_PATH    (SRC,  a8b8g8r8, null,     b5g6r5,   mmx_composite_src_x888_0565       ),
+    PIXMAN_STD_FAST_PATH    (SRC,  x8r8g8b8, null,     r5g6b5,   mmx_composite_src_x888_0565       ),
+    PIXMAN_STD_FAST_PATH    (SRC,  x8b8g8r8, null,     b5g6r5,   mmx_composite_src_x888_0565       ),
     PIXMAN_STD_FAST_PATH    (SRC,  solid,    a8,       a8r8g8b8, mmx_composite_src_n_8_8888        ),
     PIXMAN_STD_FAST_PATH    (SRC,  solid,    a8,       x8r8g8b8, mmx_composite_src_n_8_8888        ),
     PIXMAN_STD_FAST_PATH    (SRC,  solid,    a8,       a8b8g8r8, mmx_composite_src_n_8_8888        ),