SkColorCubeFilter_opts: rounding is actually free here.

src/opts/SkColorCubeFilter_opts.h

 // Copyright 2015 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #ifndef SkColorCubeFilter_opts_DEFINED
 #define SkColorCubeFilter_opts_DEFINED
 
 #include "SkColor.h"
 #include "SkNx.h"
 #include "SkUnPreMultiply.h"
@@ skipping 32 matching lines @@
         const Sk4f g0b0(g0*b0),
                    g0b1(g0*b1),
                    g1b0(g1*b0),
                    g1b1(g1*b1);
 
         const int i00 = (colorToIndex[0][g] + colorToIndex[0][b] * dim) * dim;
         const int i01 = (colorToIndex[0][g] + colorToIndex[1][b] * dim) * dim;
         const int i10 = (colorToIndex[1][g] + colorToIndex[0][b] * dim) * dim;
         const int i11 = (colorToIndex[1][g] + colorToIndex[1][b] * dim) * dim;
 
-        Sk4f color(0,0,0,0);
+        Sk4f color(0.5f);  // Starting from 0.5f gets us rounding for free.
         for (int x = 0; x < 2; ++x) {
             const int ix = colorToIndex[x][r];
 
             const SkColor lutColor00 = colorCube[ix + i00];
             const SkColor lutColor01 = colorCube[ix + i01];
             const SkColor lutColor10 = colorCube[ix + i10];
             const SkColor lutColor11 = colorCube[ix + i11];
 
             Sk4f  sum = Sk4f::FromBytes((const uint8_t*)&lutColor00) * g0b0;
             sum = sum + Sk4f::FromBytes((const uint8_t*)&lutColor01) * g0b1;
@@ skipping 11 matching lines @@
     #endif
         uint8_t* dstBytes = (uint8_t*)(dst+i);
         color.toBytes(dstBytes);
         dstBytes[SK_A32_SHIFT/8] = a;
     }
 }
 
 }  // namespace SK_OPTS NS
 
 #endif  // SkColorCubeFilter_opts_DEFINED