Improve the performance of SkConvertConfig8888Pixels using Array lookup

src/core/SkConfig8888.cpp

 #include "SkConfig8888.h"
 #include "SkMathPriv.h"
+#include "SkUnPreMultiply.h"
 
 namespace {
 
 template <int A_IDX, int R_IDX, int G_IDX, int B_IDX>
 inline uint32_t pack_config8888(uint32_t a, uint32_t r,
                                 uint32_t g, uint32_t b) {
 #ifdef SK_CPU_LENDIAN
     return (a << (A_IDX * 8)) | (r << (R_IDX * 8)) |
            (g << (G_IDX * 8)) | (b << (B_IDX * 8));
 #else
@@ skipping 37 matching lines @@
  * and an is_premul bool as params to another template function. Then IN_CFG is
  * expanded via another function call.
  */
 
 template <bool OUT_PM, int OUT_A_IDX, int OUT_R_IDX, int OUT_G_IDX, int OUT_B_IDX,
           bool IN_PM,  int IN_A_IDX,  int IN_R_IDX,  int IN_G_IDX,  int IN_B_IDX>
 inline uint32_t convert_pixel(uint32_t pixel) {
     uint32_t a, r, g, b;
     unpack_config8888<IN_A_IDX, IN_R_IDX, IN_G_IDX, IN_B_IDX>(pixel, &a, &r, &g, &b);
     if (IN_PM && !OUT_PM) {
-        // We're doing the explicit divide to match WebKit layout
-        // test expectations. We can modify and rebaseline if there
-        // it can be shown that there is a more performant way to
-        // unpremul.
+        // Using SkUnPreMultiply::ApplyScale is faster than (value * 0xff) / a.
         if (a) {
-            r = r * 0xff / a;
-            g = g * 0xff / a;
-            b = b * 0xff / a;
+            SkUnPreMultiply::Scale scale = SkUnPreMultiply::GetScale(a);
+            r = SkUnPreMultiply::ApplyScale(scale, r);
+            g = SkUnPreMultiply::ApplyScale(scale, g);
+            b = SkUnPreMultiply::ApplyScale(scale, b);
         } else {
             return 0;
         }
     } else if (!IN_PM && OUT_PM) {
-        // This matches WebKit's conversion which we are replacing.
-        // We can consider alternative rounding rules for performance.
-        r = SkMulDiv255Ceiling(r, a);
-        g = SkMulDiv255Ceiling(g, a);
-        b = SkMulDiv255Ceiling(b, a);
+        // This matches SkUnPreMultiply conversion which we are replacing.
+        r = SkMulDiv255Round(r, a);
+        g = SkMulDiv255Round(g, a);
+        b = SkMulDiv255Round(b, a);
     }
     return pack_config8888<OUT_A_IDX, OUT_R_IDX, OUT_G_IDX, OUT_B_IDX>(a, r, g, b);
 }
 
 template <bool OUT_PM, int OUT_A_IDX, int OUT_R_IDX, int OUT_G_IDX, int OUT_B_IDX, SkCanvas::Config8888 IN_CFG>
 inline uint32_t convert_pixel(uint32_t pixel) {
     switch(IN_CFG) {
         case SkCanvas::kNative_Premul_Config8888:
             return convert_pixel<OUT_PM, OUT_A_IDX,       OUT_R_IDX,       OUT_G_IDX,       OUT_B_IDX,
                                  true,  SK_NATIVE_A_IDX,  SK_NATIVE_R_IDX, SK_NATIVE_G_IDX, SK_NATIVE_B_IDX>(pixel);
@@ skipping 186 matching lines @@
         case SkCanvas::kBGRA_Unpremul_Config8888:
             return pack_config8888<3, 2, 1, 0>(a, r, g, b);
         case SkCanvas::kRGBA_Premul_Config8888:
         case SkCanvas::kRGBA_Unpremul_Config8888:
             return pack_config8888<3, 0, 1, 2>(a, r, g, b);
         default:
             SkDEBUGFAIL("Unexpected config8888");
             return 0;
     }
 }