mmx: add and use pack_4x565 function

The pack_4x565 makes use of the pack_4xpacked565 function which uses pmadd.

Some of the speed up is probably attributable to removing the artificial
serialization imposed by the
	vdest = pack_565 (..., vdest, 0);
	vdest = pack_565 (..., vdest, 1);
	...
pattern.

Loongson:
        over_n_0565 =  L1:  16.44  L2:  16.42  M: 13.83 (  9.85%)  HT: 12.83  VT: 12.61  R: 12.34  RT:  8.90 (  93Kops/s)
        over_n_0565 =  L1:  42.48  L2:  42.53  M: 29.83 ( 21.20%)  HT: 23.39  VT: 23.72  R: 21.80  RT: 11.60 ( 113Kops/s)

     over_8888_0565 =  L1:  15.61  L2:  15.42  M: 12.11 ( 25.79%)  HT: 11.07  VT: 10.70  R: 10.37  RT:  7.25 (  82Kops/s)
     over_8888_0565 =  L1:  35.01  L2:  35.20  M: 21.42 ( 45.57%)  HT: 18.12  VT: 17.61  R: 16.09  RT:  9.01 (  97Kops/s)

      over_n_8_0565 =  L1:  15.17  L2:  14.94  M: 12.57 ( 17.86%)  HT: 11.96  VT: 11.52  R: 10.79  RT:  7.31 (  79Kops/s)
      over_n_8_0565 =  L1:  29.83  L2:  29.79  M: 21.85 ( 30.94%)  HT: 18.82  VT: 18.25  R: 16.15  RT:  8.72 (  91Kops/s)

over_n_8888_0565_ca =  L1:  15.25  L2:  15.02  M: 11.64 ( 41.39%)  HT: 11.08  VT: 10.72  R: 10.02  RT:  7.00 (  77Kops/s)
over_n_8888_0565_ca =  L1:  30.12  L2:  29.99  M: 19.47 ( 68.99%)  HT: 17.05  VT: 16.55  R: 14.67  RT:  8.38 (  88Kops/s)

ARM/iwMMXt:
        over_n_0565 =  L1:  19.29  L2:  19.88  M: 17.38 ( 10.54%)  HT: 15.53  VT: 16.11  R: 13.69  RT: 11.00 (  96Kops/s)
        over_n_0565 =  L1:  36.02  L2:  34.85  M: 28.04 ( 16.97%)  HT: 22.12  VT: 24.21  R: 22.36  RT: 12.22 ( 103Kops/s)

     over_8888_0565 =  L1:  18.38  L2:  16.59  M: 12.34 ( 22.29%)  HT: 11.67  VT: 11.71  R: 11.02  RT:  6.89 (  72Kops/s)
     over_8888_0565 =  L1:  24.96  L2:  22.17  M: 15.11 ( 26.81%)  HT: 14.14  VT: 13.71  R: 13.18  RT:  8.13 (  78Kops/s)

      over_n_8_0565 =  L1:  14.65  L2:  12.44  M: 11.56 ( 14.50%)  HT: 10.93  VT: 10.39  R: 10.06  RT:  7.05 (  70Kops/s)
      over_n_8_0565 =  L1:  18.37  L2:  14.98  M: 13.97 ( 16.51%)  HT: 12.67  VT: 10.35  R: 11.80  RT:  8.14 (  74Kops/s)

over_n_8888_0565_ca =  L1:  14.27  L2:  12.93  M: 10.52 ( 33.23%)  HT:  9.70  VT:  9.90  R:  9.31  RT:  6.34 (  65Kops/s)
over_n_8888_0565_ca =  L1:  19.69  L2:  17.58  M: 13.40 ( 42.35%)  HT: 11.75  VT: 11.33  R: 11.17  RT:  7.49 (  73Kops/s)

1 files changed, 52 insertions, 55 deletions

diff --git a/pixman/pixman-mmx.c b/pixman/pixman-mmx.c
index b14201a4..01a2bc93 100644
--- a/pixman/pixman-mmx.c
+++ b/pixman/pixman-mmx.c
@@ -598,6 +598,12 @@ pack_4xpacked565 (__m64 a, __m64 b)
 #endif
 }
 
+static force_inline __m64
+pack_4x565 (__m64 v0, __m64 v1, __m64 v2, __m64 v3)
+{
+    return pack_4xpacked565 (pack8888 (v0, v1), pack8888 (v2, v3));
+}
+
 #ifndef _MSC_VER
 
 static force_inline __m64
@@ -1396,16 +1402,14 @@ mmx_composite_over_n_0565 (pixman_implementation_t *imp,
 
 	while (w >= 4)
 	{
-	    __m64 vdest;
+	    __m64 vdest = *(__m64 *)dst;
 
-	    vdest = *(__m64 *)dst;
-
-	    vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 0)), vdest, 0);
-	    vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 1)), vdest, 1);
-	    vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 2)), vdest, 2);
-	    vdest = pack_565 (over (vsrc, vsrca, expand565 (vdest, 3)), vdest, 3);
+	    __m64 v0 = over (vsrc, vsrca, expand565 (vdest, 0));
+	    __m64 v1 = over (vsrc, vsrca, expand565 (vdest, 1));
+	    __m64 v2 = over (vsrc, vsrca, expand565 (vdest, 2));
+	    __m64 v3 = over (vsrc, vsrca, expand565 (vdest, 3));
 
-	    *(__m64 *)dst = vdest;
+	    *(__m64 *)dst = pack_4x565 (v0, v1, v2, v3);
 
 	    dst += 4;
 	    w -= 4;
@@ -1818,22 +1822,19 @@ mmx_composite_over_8888_0565 (pixman_implementation_t *imp,
 
 	while (w >= 4)
 	{
-	    __m64 vsrc0, vsrc1, vsrc2, vsrc3;
-	    __m64 vdest;
+	    __m64 vdest = *(__m64 *)dst;
 
-	    vsrc0 = load8888 ((src + 0));
-	    vsrc1 = load8888 ((src + 1));
-	    vsrc2 = load8888 ((src + 2));
-	    vsrc3 = load8888 ((src + 3));
+	    __m64 vsrc0 = load8888 ((src + 0));
+	    __m64 vsrc1 = load8888 ((src + 1));
+	    __m64 vsrc2 = load8888 ((src + 2));
+	    __m64 vsrc3 = load8888 ((src + 3));
 
-	    vdest = *(__m64 *)dst;
-
-	    vdest = pack_565 (over (vsrc0, expand_alpha (vsrc0), expand565 (vdest, 0)), vdest, 0);
-	    vdest = pack_565 (over (vsrc1, expand_alpha (vsrc1), expand565 (vdest, 1)), vdest, 1);
-	    vdest = pack_565 (over (vsrc2, expand_alpha (vsrc2), expand565 (vdest, 2)), vdest, 2);
-	    vdest = pack_565 (over (vsrc3, expand_alpha (vsrc3), expand565 (vdest, 3)), vdest, 3);
+	    __m64 v0 = over (vsrc0, expand_alpha (vsrc0), expand565 (vdest, 0));
+	    __m64 v1 = over (vsrc1, expand_alpha (vsrc1), expand565 (vdest, 1));
+	    __m64 v2 = over (vsrc2, expand_alpha (vsrc2), expand565 (vdest, 2));
+	    __m64 v3 = over (vsrc3, expand_alpha (vsrc3), expand565 (vdest, 3));
 
-	    *(__m64 *)dst = vdest;
+	    *(__m64 *)dst = pack_4x565 (v0, v1, v2, v3);
 
 	    w -= 4;
 	    dst += 4;
@@ -2368,25 +2369,22 @@ mmx_composite_over_n_8_0565 (pixman_implementation_t *imp,
 	    }
 	    else if (m0 | m1 | m2 | m3)
 	    {
-		__m64 vdest;
-		__m64 vm0, vm1, vm2, vm3;
-
-		vdest = *(__m64 *)dst;
+		__m64 vdest = *(__m64 *)dst;
 
-		vm0 = to_m64 (m0);
-		vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm0),
-					   expand565 (vdest, 0)), vdest, 0);
-		vm1 = to_m64 (m1);
-		vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm1),
-					   expand565 (vdest, 1)), vdest, 1);
-		vm2 = to_m64 (m2);
-		vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm2),
-					   expand565 (vdest, 2)), vdest, 2);
-		vm3 = to_m64 (m3);
-		vdest = pack_565 (in_over (vsrc, vsrca, expand_alpha_rev (vm3),
-					   expand565 (vdest, 3)), vdest, 3);
-
-		*(__m64 *)dst = vdest;
+		__m64 vm0 = to_m64 (m0);
+		__m64 v0 = in_over (vsrc, vsrca, expand_alpha_rev (vm0),
+					   expand565 (vdest, 0));
+		__m64 vm1 = to_m64 (m1);
+		__m64 v1 = in_over (vsrc, vsrca, expand_alpha_rev (vm1),
+					   expand565 (vdest, 1));
+		__m64 vm2 = to_m64 (m2);
+		__m64 v2 = in_over (vsrc, vsrca, expand_alpha_rev (vm2),
+					   expand565 (vdest, 2));
+		__m64 vm3 = to_m64 (m3);
+		__m64 v3 = in_over (vsrc, vsrca, expand_alpha_rev (vm3),
+					   expand565 (vdest, 3));
+
+		*(__m64 *)dst = pack_4x565 (v0, v1, v2, v3);;
 	    }
 
 	    w -= 4;
@@ -2483,24 +2481,23 @@ mmx_composite_over_pixbuf_0565 (pixman_implementation_t *imp,
 
 	    if ((a0 & a1 & a2 & a3) == 0xFF)
 	    {
-		__m64 vdest;
-		vdest = pack_565 (invert_colors (load8888 (&s0)), _mm_setzero_si64 (), 0);
-		vdest = pack_565 (invert_colors (load8888 (&s1)), vdest, 1);
-		vdest = pack_565 (invert_colors (load8888 (&s2)), vdest, 2);
-		vdest = pack_565 (invert_colors (load8888 (&s3)), vdest, 3);
+		__m64 v0 = invert_colors (load8888 (&s0));
+		__m64 v1 = invert_colors (load8888 (&s1));
+		__m64 v2 = invert_colors (load8888 (&s2));
+		__m64 v3 = invert_colors (load8888 (&s3));
 
-		*(__m64 *)dst = vdest;
+		*(__m64 *)dst = pack_4x565 (v0, v1, v2, v3);
 	    }
 	    else if (s0 | s1 | s2 | s3)
 	    {
 		__m64 vdest = *(__m64 *)dst;
 
-		vdest = pack_565 (over_rev_non_pre (load8888 (&s0), expand565 (vdest, 0)), vdest, 0);
-		vdest = pack_565 (over_rev_non_pre (load8888 (&s1), expand565 (vdest, 1)), vdest, 1);
-		vdest = pack_565 (over_rev_non_pre (load8888 (&s2), expand565 (vdest, 2)), vdest, 2);
-		vdest = pack_565 (over_rev_non_pre (load8888 (&s3), expand565 (vdest, 3)), vdest, 3);
+		__m64 v0 = over_rev_non_pre (load8888 (&s0), expand565 (vdest, 0));
+		__m64 v1 = over_rev_non_pre (load8888 (&s1), expand565 (vdest, 1));
+		__m64 v2 = over_rev_non_pre (load8888 (&s2), expand565 (vdest, 2));
+		__m64 v3 = over_rev_non_pre (load8888 (&s3), expand565 (vdest, 3));
 
-		*(__m64 *)dst = vdest;
+		*(__m64 *)dst = pack_4x565 (v0, v1, v2, v3);
 	    }
 
 	    w -= 4;
@@ -2675,12 +2672,12 @@ mmx_composite_over_n_8888_0565_ca (pixman_implementation_t *imp,
 	    {
 		__m64 vdest = *(__m64 *)q;
 
-		vdest = pack_565 (in_over (vsrc, vsrca, load8888 (&m0), expand565 (vdest, 0)), vdest, 0);
-		vdest = pack_565 (in_over (vsrc, vsrca, load8888 (&m1), expand565 (vdest, 1)), vdest, 1);
-		vdest = pack_565 (in_over (vsrc, vsrca, load8888 (&m2), expand565 (vdest, 2)), vdest, 2);
-		vdest = pack_565 (in_over (vsrc, vsrca, load8888 (&m3), expand565 (vdest, 3)), vdest, 3);
+		__m64 v0 = in_over (vsrc, vsrca, load8888 (&m0), expand565 (vdest, 0));
+		__m64 v1 = in_over (vsrc, vsrca, load8888 (&m1), expand565 (vdest, 1));
+		__m64 v2 = in_over (vsrc, vsrca, load8888 (&m2), expand565 (vdest, 2));
+		__m64 v3 = in_over (vsrc, vsrca, load8888 (&m3), expand565 (vdest, 3));
 
-		*(__m64 *)q = vdest;
+		*(__m64 *)q = pack_4x565 (v0, v1, v2, v3);
 	    }
 	    twidth -= 4;
 	    p += 4;